Aging Effects on Task-Evoked Pupillary Responses By Mingjian He Sc.B. Brown University, 2017 Thesis Submitted in partial fulfillment of the requirements for the Degree of Master of Science in the Department of Cognitive, Linguistic, and Psychological Sciences at Brown University PROVIDENCE, RHODE ISLAND May 2018 This thesis by Mingjian He is accepted in its present form by the Department of Cognitive, Linguistic, and Psychological Sciences as satisfying the thesis requirement for the degree of Master of Science Date _____________ ________________________________ William C. Heindel, Advisor Approved by the Graduate Council Date _____________ ________________________________ Andrew G. Campbell, Dean of the Graduate school ii Table of Contents INTRODUCTION .................................................................................................................. 1 METHOD ............................................................................................................................. 14 RESULTS ............................................................................................................................. 22 DISCUSSION ........................................................................................................................ 32 REFERENCE ....................................................................................................................... 53 TABLES ................................................................................................................................ 61 FIGURES .............................................................................................................................. 63 iii Introduction Noradrenaline (NA), or norepinephrine, is one of the classical neurotransmitters in the brain. NA is mainly produced by neurons of locus coeruleus (LC) nuclei in the brainstem (Fuxe, 1965), and it gets broadly projected over the entire neocortex to modulate cortical processes. As a neuromodulatory network, the LC-NA system is critical for regulating normal arousal states of the brain (Aston-Jones & Cohen, 2005). Arousal can be described in more biological terms as the facilitation of neuronal processing of relevant information. Information processing mediated by NA is fundamental to a wide range of cortical and subcortical circuits for sensory, attention, and memory functions (Berridge & Waterhouse, 2003). Therefore, it is not surprising that the LC-NA system has pervasive influence on cognitive and behavioral functioning. Normal LC-NA signaling might be adversely impacted by aging, and changes in the LC-NA system can contribute to age-related cognitive deficits. It has long been reported that there is age-related reduction in the NA concentration (Spokes, 1979). A review article reported progressive decrease in NA concentrations in several areas of the brain with age, with more than 25% of the original cells in LC nuclei lost by the age of 90 (Mann, 1983). By the age of 63, the neuronal population of locus-coeruleus and sub- coeruleus nuclei can be reduced by 40% (Vijayashankar & Brody, 1979). Reduction in NA neuromodulation has been hypothesized to be a key contributor to age-related cognitive decline during normal aging (Arnsten & Goldman-Rakic, 1985). A recent large-scale autopsy study related LC neuronal density to late-life cognition and highlighted the importance of the LC-NA system in maintaining normal cognition in the 1 presence of age-related or Alzheimer’s disease-related pathologies (Wilson et al., 2013). Indeed, the LC-NA system is consistently associated with cognitive functioning in various domains, including attentional alerting, memory performance, and conflict resolution. Each of these three cognitive domains and their connections to the LC-NA system in aging will be briefly summarized below. The alerting network, as one of the three components of the attention system (Posner & Petersen, 1990), is perhaps the most direct translation of brain arousal states to an explicit cognitive process with behavioral consequences. Presentation of non- informative warning cues prior to task stimuli can elicit faster overt responses (Bernstein, Rose, & Ashe, 1970; Hackley & Valle-Inclan, 1998; Posner, Klein, Summers, & Buggie, 1973), and this behavioral effect has been attributed to a phasic increase in the arousal level triggered by cues, which enables faster target detections and rapid response actions (Bernstein, Chu, Briggs, & Schurman, 1973; Kristjansson, Stern, Brown, & Rohrbaugh, 2009; Petersen & Posner, 2012; Sanders, 1980). Specifically, this phasic increase in arousal is hypothesized to be neuronally implemented as an increased noradrenergic innervation from LC nuclei to cortical sensory processing areas (Aston-Jones & Cohen, 2005; Howells, Stein, & Russell, 2012). Consistent with the pattern of reduced LC-NA signaling with aging, in arousal-response studies using variants of the Posner Cueing Task (Posner et al., 1973), elderly participants exhibited reduced behavioral alerting effect elicited by non-informative warning cues (Festa, Ott, & Heindel, 2004; Jennings, Dagenbach, Engle, & Funke, 2007; Tales, Muir, Bayer, & Snowden, 2002). Although there is lack of direct evidence, it is reasonable to suspect that compromised functioning of the LC-NA system might underlie these observed behavioral deficits. 2 Memory functioning is another cognitive domain where the LC-NA system plays an important neuromodulatory role. Since the early proposition of the involvement of noradrenergic projections in cortical learning (Kety, 1972), increased NA signaling is now thought to facilitate signal-to-noise ratio by reducing spontaneous activities of sensory neurons and enhancing synchronous spiking (Bouret & Sara, 2002; Lecas, 2004). During memory retrieval, stimulation of LC neurons can alleviate amnesia in rodent models (Devauges & Sara, 1991). Furthermore, facilitation of retrieval of previous learnt memory by context cues is dependent on the activation of the LC-based arousal system (Sara, 2000; Sara & Devauges, 1988; Sara & Hars, 2006). These findings are supported by functional imaging studies in humans highlighting the LC-amygdala circuit that activates fronto-hippocampal networks during memory retrieval (Sterpenich et al., 2006). It is clear that normal aging is accompanied by declines in memory functioning (Buckner, 2004; Harada, Love, & Triebel, 2013). A study on rodent models demonstrated reduced level of NA concentration accompanied by age-related impairment in emotional memory, which could be alleviated through pharmacological interventions to increase NA signaling (Luo et al., 2015). Therefore, it would be revealing to examine the generalization of similar findings to human subjects in order to understand whether changes in LC-NA activities might mediate memory declines in normal aging. As a signature PFC function, conflict resolution is mediated by extensive fronto- striatal networks, and normal cognitive control is sensitive to small changes in multiple neurotransmitter systems including acetylcholine, dopamine, and noradrenaline (Arnsten & Rubia, 2012). Traditionally, processing of a conflict signal during decision-making is thought to be accomplished in an extended network involving the anterior cingulate 3 cortex (ACC) region (Botvinick, Cohen, & Carter, 2004; MacDonald, Cohen, Stenger, & Carter, 2000). While research on PFC functioning often focus on dopamine, NA is also studied as another catecholamine with pervasive impact on working memory and other PFC functions (Arnsten, 1993; Li, Lindenberger, & Sikström, 2001). LC-NA signaling has been hypothesized to play a mediating role in conflict resolution and exertion of cognitive control. The Adaptation By Binding model suggests that conflict detection would elicit a phasic arousal response, which serves as a broadcasting signal of the demand of cognitive control (Abrahamse, Braem, Notebaert, & Verguts, 2016). In addition, NA signaling has been suggested to maintain attentional focus on the target stimulus when there are distractors present in the same visual field (Tiplady, Bowness, Stien, & Drummond, 2005), which might be computationally similar to the enhancement of signal-to-noise ratio in memory functioning. More generally, the LC-NA system might modulate synaptic strengths to facilitate fronto-parietal and LC-intraparietal sulcus connectivities to execute cognitive control in cortical sensory areas (Coull, Büchel, Friston, & Frith, 1999). Normal aging has been shown to impact cognitive control capacities (Braver & Barch, 2002), which is a deficit associated with reduced functional connectivity in the cognitive control networks (Baddeley, Baddeley, Bucks, & Wilcock, 2001; Campbell, Grady, Ng, & Hasher, 2012). Indeed, aging population often exhibited impairments with inhibitory responses in conflict processing tasks (Castel, Balota, Hutchison, Logan, & Yap, 2007). Despite predominant connections of these behavioral patterns to dopaminergic changes with aging (Braver & Barch, 2002), it is of interest to understand whether alterations in the LC-NA system are also involved in age-related impairments of conflict resolution. 4 One practical difficulty encountered by research on the relationship between the LC-NA system and cognition during aging is to measure LC integrity. Previous literature demonstrating age-related decline in LC neuronal density and NA concentration were based on post-mortem brain tissue samples. Not only little is known about the topographical distribution of cell loss in LC over time (Manaye, McIntire, Mann, & German, 1995; Marcyniuk, Mann, & Yates, 1989), it is also unclear whether reduction in the number of LC neurons would linearly translate into cognitive and behavioral impairments. Furthermore, the claim itself of decline in LC cell counts with aging has been challenged. Some recent neurobiological studies have reported preserved absolute number of neurons in aged non-demented persons (Kubis et al., 2000; Mouton, Pakkenberg, Gundersen, & Price, 1994). Given the conflicting results among cell- stereology studies, a recent influential review concluded that it is uncertain whether LC cell counts decrease with age, but age-related changes in the LC-NA system likely modulate normal cognitive abilities in the aging population (Mather & Harley, 2016). Implicit in this statement is moving away from assessing LC integrity using cell biological metrics toward emphasizing on the involvement of LC-NA signaling in real time cognitive performance. It is a reasonable shift of experimental approach because while it is important to document the decline in LC density, the net output from LC nuclei could remain stable due to increased innervation from the remaining neurons, until a substantial amount of cell population has been lost (Hoogendijk et al., 1999; Miller, Kolb, Leverenz, Peskind, & Raskind, 1999). Ultimately, it is the functional relevance of LC-NA signaling most important for understanding its role in mediating age-related cognitive declines. Hence, it becomes critical to assess task-evoked LC-NA responses in 5 human subjects. Unfortunately, it is just as difficult as estimating LC cell counts to obtain in-vivo neural signals from the LC-NA system during task performance. This difficulty largely stems from the small size of LC nuclei and their location being buried deep in the brainstem (Schwarz & Luo, 2015). One possible solution is provided in the pupillometry literature and the use of eye-tracking techniques during cognitive tasks. Since early work showing correlations between pupil diameter and problem- solving difficulty (Hess & Polt, 1963), the use of pupillometry has expanded greatly in the field of cognitive science (Sirois & Brisson, 2014). Despite the application of eye- tracking in different task paradigms, pupil dilation (PD) has been consistently linked to the arousal system (Eckstein, Guerra-Carrillo, Singley, & Bunge, 2017). Increased arousal levels in tasks involving attentional allocation, emotional stimuli, and difficult task demand are also reflected in increased pupil diameter (McDougal & Gamlin, 2015; Samuels & Szabadi, 2008). These findings can be loosely summarized in the statement that pupil diameter is modulated by sympathetic and parasympathetic activities (Breen, Burde, & Loewy, 1983). Critically, electrophysiological studies on monkeys showed that stimulation of LC neurons can elicit a parallel time-locked increase in pupil diameter (Rajkowski, Kubiak, & Aston-Jones, 1994; Samuels & Szabadi, 2008). This finding suggests a tight link between PD responses and LC-NA signaling. Recent studies recording from a larger population of neurons replicated this finding and demonstrated that phasic bursts of LC neurons and adjacent nuclei are accompanied by robust increases in pupil diameter (Joshi, Li, Kalwani, & Gold, 2016; Reimer et al., 2016). A human fMRI study also provided some preliminary evidence of the feasibility of relating PD responses to BOLD signals (Murphy, O’Connell, O’Sullivan, Robertson, & Balsters, 2014). In 6 general, reasonably strong evidence suggests that greater PD can indirectly index greater LC-NA activities. It should be noted that despite similar application of eye-tracking in different task paradigms, pupillary responses are far from a homogeneous response. In other words, specific mechanistic accounts of arousal and the involvement of NA signaling in task-related cognitive processes do vary from task to task (Gilzenrat, Nieuwenhuis, & Cohen, 2010; Murphy, Robertson, Balsters, & O’connell, 2011; Naber, Frässle, Rutishauser, & Einhäuser, 2013; Nassar et al., 2012). Nevertheless, pupillary responses in different task paradigms might be jointly understood as an arousal response at the cognitive level (Aston-Jones & Cohen, 2005) and an alteration in LC-NA signaling at the neurobiological level (Sara & Bouret, 2012). PD can therefore be viewed as a psychophysiological reporter variable for overall arousal level (Beatty & Lucero-Wagoner, 2000). Under the assumption that PD is associated with LC-NA signaling, PD can be further treated as a proxy measure of LC- NA activities. In addition to the aforementioned correlational studies, this assumption is substantiated by some proposals at the neuronal level suggesting that a phasic burst of NA might bind to inhibitory α2-adrenoceptors in the parasympathetic Edinger-Westphal nucleus, which results in the relaxation of the sphincter muscles coupled with sympathetically innervated dilatory muscles, and therefore PD ensues (Breen et al., 1983; Elam, Thorén, & Svensson, 1986; Nieuwenhuis, De Geus, & Aston-Jones, 2011). With the rationale of using PD to assess the neuromodulatory role of the LC-NA system in cognitive tasks, PD has been since associated with a wide range of cognitive processes during task performance. A number of extensive reviews have surveyed the use of pupillary responses in different cognitive domains. First, amygdala-based emotional 7 processing of negative valence has been found to increase PD (Granholm & Steinhauer, 2004); second, increased cognitive load and task difficulty have been shown to increase physiological arousal and elicit increased PD (Just, Carpenter, & Miyake, 2003); finally, PD has also been used as an independent assessment of cost of cognitive control, with greater exertion of cognitive control leading to greater PD (van der Wel & van Steenbergen, 2018). Mirroring the order of previous discussion on the attentional alerting, memory performance, and conflict resolution cognitive domains, the following paragraphs will briefly summarize available literature on age-related changes in pupillary responses in each of the three domains. In the arousal literature, pupillometry studies on healthy young adults identified increased PD to non-informative warning cues using auditory tones (Gabay, Pertzov, & Henik, 2011; Geva, Zivan, Warsha, & Olchik, 2013; Tona, Murphy, Brown, & Nieuwenhuis, 2016). This reported increase in PD to warning cues is consistent with the idea of a phasic burst of LC-NA signaling in response to alerting stimuli, reflecting an increased arousal level. However, inconsistent results have been reported on pupillary responses in the aging population. One study showed comparable amplitudes of pupillary responses between age groups (Kim, Beversdorf, & Heilman, 2000). But in another study, elderly participants were shown to produce attenuated pupillary arousal responses to processing of emotional faces (Allard, Wadlinger, & Isaacowitz, 2010). Therefore, it is unknown whether task-evoked LC-NA signaling in phasic alerting tasks as indexed by PD reliably change with normal aging. 8 LC-NA activities in memory retrieval have been operationally defined in pupillometry studies in terms of a pupillary old/new effect. A substantial amount of studies consistently identified increased PD for old items compared to new items, hence the naming of old/new effect, during memory retrieval in young adults (Brocher & Graf, 2017; Goldinger & Papesh, 2012; Heaver & Hutton, 2011; Kafkas & Montaldi, 2015). It has been suggested that retrieving old memory might be more effortful than judging new items, resulting in an increased pupillary response (Võ et al., 2008). An alternative account argued that phasic PD responses reflect the greater memory traces of old items compared to new items (Otero, Weekes, & Hutton, 2011; Papesh, Goldinger, & Hout, 2012). The second theory is more congruent with neurobiological explanations of the involvement of NA in memory retrieval. Only one existing study examined the aging effect on this pupillary old/new effect, and comparable PD responses to old stimuli were reported in the two age groups, suggesting a similar response across age groups to difference in memory strengths of old and new stimuli (Hämmerer et al., 2017). Two additional studies of the aging population combined eye-tracking with tasks including a recognition memory component, but they did not report on this particular difference between old and new items in terms of pupil dilatory responses (Dragan et al., 2017; Ziaei, von Hippel, Henry, & Becker, 2015). Lastly, previous studies have identified PD responses in tasks engaging cognitive control in young adults (Schacht, Dimigen, & Sommer, 2010; Scharinger, Soutschek, Schubert, & Gerjets, 2015; van Bochove, Van der Haegen, Notebaert, & Verguts, 2013; van Steenbergen & Band, 2013). Simon Interference Task (Simon & Wolf, 1963) and the Flanker Task (Eriksen & Eriksen, 1974) are commonly employed to introduce decision 9 conflicts. In this specific form of conflict processing, incongruent trials elicited increased PD compared to congruent trials, reflecting an increased need for conflict resolution and elevated processing demand. Identification of increased PD in these tasks is again consistent with the purported role LC-NA signaling in mediating and executing cognitive control during decision-making and response selection. Direct evidence of aging effects on LC-NA signaling in conflict resolution as indexed by PD is missing. Related studies looking at the influence of cognitive load manipulations and speech comprehension difficulties have provided some, despite conflicting, clues. One study reported reduced pupillary responses trigged by increased memory load for elderly participants in a Sternberg memory-search task (Van Gerven, Paas, Van Merriënboer, & Schmidt, 2004). However, another study reported the opposite pattern, showing increased pupillary responses in elderly participants compared to young adults when task difficulties was increased in a sentence recall task. The authors interpreted this pattern as reflecting greater cognitive demand imposed by the same difficulty manipulation for the elderly participants (Piquado, Isaacowitz, & Wingfield, 2010). Hence, it remains inconclusive whether there is age-related reduction in pupillary responses reflecting compromised LC- NA signaling in the cognitive control domain. Inconsistency in reports of aging effects on pupillary responses is at least in part due to complications with measuring PD in elderly participants. The first problem comes from that light intensity can substantially modulate pupil diameter in cognitive tasks (Bradshaw, 1969; Gilzenrat et al., 2010). Visuopercetual tasks often unavoidably change luminance levels when presenting task-relevant stimuli. Showing a bright stimulus on a dark background would cause pupil constriction; conversely, appearance of a dark 10 stimulus on a white background would lead to dilation. A major concern is that these light reflexes could obscure the task-evoked pupillary responses measured by monitoring pupil diameter. Indeed, existing literature has repeatedly reported the influence of ambient luminance on pupillary responses (Peysakhovich, Vachon, & Dehais, 2017). This issue is especially problematic if one wishes to use PD responses as a proxy measure of LC-NA activities to understand age-related changes in cognitive functioning (Moloney et al., 2006), because aging has been shown to impact the magnitude of pupil constriction to visual stimuli (Bergamin & Kardon, 2002; Loewenfeld & Lowenstein, 1993). Thus, it becomes uncertain whether some studies found reduced pupillary responses in older adults compared to young controls simply as results of weakened peripheral muscular structures with age and a smaller range allowed for potential PD during task performance (Van Gerven et al., 2004). A separate issue arises from the analytic approach to pupillary data taken by previous studies. In the traditional method, average baseline pupil dimeter is subtracted from trial pupil diameter during recorded time windows to yield dilatory pupil sizes. Arbitrary time periods are then defined to average PD to compute pupillary response measures (Beatty & Lucero-Wagoner, 2000). Subtraction of baseline could introduce noise in measures used in analyses and result in statistical dependence between baseline pupil diameter and trial PD (Murphy et al., 2011). Some evidence suggested that baseline pupil diameter provides information regarding sustained attention (Reimer et al., 2014; Unsworth & Robison, 2016; Van Den Brink, Murphy, & Nieuwenhuis, 2016); therefore, calculating task-evoked pupillary responses by subtracting baseline measures might confound changes in phasic arousal with changes in tonic arousal levels. In addition, 11 averaging trial PD across time windows assumes homogeneity of pupillary responses across multiple time points. Yet, time windows used for averaging are often selected subjectively, and they might bias test statistics in unjustified ways (Sirois & Brisson, 2014). Hence, a more data-driven approach should be adopted to characterize pupillary responses evoked by task manipulations in cognitive tasks. Based on the above discussion, this study aimed at characterizing aging effects on task-evoked pupillary responses in multiple cognitive domains to explore functional changes in phasic responses of the LC-NA system during task performance. While the heterogeneity of pupillary responses in different task paradigms were well recognized, a secondary goal in this study was to identify general patterns in age-related changes in the LC-NA system cross-cutting distinct cognitive domains, as the same subjects underwent all three cognitive tasks. Specifically, the attentional alerting effect of a non-informative auditory warning cue was assessed in a simple target-localization task. This task served as a canonical case of investigating PD-indexed changes in arousal levels and their impacts on behavioral performance. Then, pupillary old/new effects were characterized in a standard recognition memory task to test whether changes in task-evoked LC-NA signaling underlie impaired memory performance in the aging population compared to young adults. Finally, an arrow-version flanker task was administered to extend the examination of aging effects on LC-NA signaling to the conflict processing domain. These three tasks were chosen carefully to ensure central fixation can be achieved throughout tasks to minimize eye movement artifacts in measuring PD. Furthermore, all tasks must be compatible with isoluminant stimuli to avoid confounds of light reflexes. Lastly, a novel multiple regression approach to pupillary data was taken to tease apart 12 changes in baseline pupil diameter from changes in task-evoked PD, which was assumed to indirectly index LC-NA signaling. The use of linear regression allowed characterizing pupillary responses beyond those triggered by primary task manipulations. Associations between PD and behavioral response speed and interactions with task conditions were also examined to provide a much more elaborate description of task-evoked LC-NA signaling in cognitive task performance. 13 Method Participants Twenty-eight healthy young controls (YC) who were Brown University students and thirty elderly participants (ED) recruited from local communities completed all tasks in this study. Data from three YC individuals were excluded due to technical issues with the eye-tracking device. On the basis of poor neuropsychological testing results, data from four ED individuals were excluded due to scores below 27 on Mini-Mental State Examination (MMSE) and/or normalized scale index on Repeatable Battery for the Assessment of Neuropsychological Status (RBANS) being below 90 on the total scaled index. Data from one additional ED individual were excluded due to incomplete neuropsychological testing. After exclusion, twenty-five subjects were included in each age group (N = 50) for all analyses. All individuals had normal hearing based on self-report and normal or corrected-to-normal vision examined using the Tumbling-E Visual Acuity Test and the Mars Letter Contrast Sensitivity Test. All individuals had no active or recent history of psychological or neurological disorders and received monetary payment for participation. Table 1 shows the demographic and neuropsychological testing information collected for both YC and ED groups. Overall, the elderly participants were cognitively healthy. Age-related changes in neuropsychological testing were identified primarily in tasks with working memory demands but not in tasks relying on visuoperceptual abilities, indicating high functioning and preserved general cognition in the ED group. 14 Eye-tracking Apparatus Stimuli were displayed on a 19-inch CRT monitor (1152x864 pixel resolution, 72Hz, Viewsonic G90fb) situated 75cm from participants’ eyes. The tasks were delivered and controlled on an Intel Core 2 Quad 2.836GHz PC (OS Windows XP SP3 32-bit, 2002) with Nvidia GeForce GT440 graphics card and onboard RealTek/Azalia Audio soundcard using E-Prime V2.0 (Psychology Software Tools, Pittsburgh, PA). Pupil diameters and eye positions were recorded using EyeLink®1000 video- based desktop- mounted eye tracker (EyeLink1000, SR Research, Ontario, Canada) with a sampling rate of 250Hz. Participants had their heads stabilized in a chin rest, and pupil diameters from the left eye were captured using the Centroid Pupil Tracking Algorithm (EyeLink1000) by an infrared- sensitive camera. The system was calibrated and validated to < 1° average visual angle Cartesian prediction error at the onset of each block using a 9-point grid calibration. In all tasks, participants were instructed to maintain fixation at the center of the screen and minimize blinking. Throughout the tasks, blinking was encouraged to be limited to designated blink screens between trials that were outside of the 2s trial pupil measurement time windows. Isoluminant Task Stimuli For all tasks, an isoluminant color-scheme was created by using modified colors from the Teufel colors set (Teufel & Wehrhahn, 2000). A slate blue (RGB = [55, 123, 170], CIE-xyY = [0.196, 0.199, 25.5]) background and dark orange (RGB = [186, 94, 47], CIE-xyY = [0.234, 0.226, 25.8]) task stimuli were used to maintain a constant light 15 intensity at 25cd/m2 throughout all tasks, verified with ColorCAL MKII Colorimeter (Cambridge Research Systems, Rochester, UK). Data Processing and Analyses Behavioral and pupillary data were pre-processed and analyzed using MATLAB R2017a (The MathWorks, Natick, MA). Artifacts and blinks were removed from subsequent analyses. To facilitate comparisons of aging effects on pupillary responses across tasks, a uniform eye-tracking scheme was determined based on PD patterns in the literature, and the scheme was applied to all tasks investigated in this study. Baseline pupil diameter was measured using the average of pupil sizes during 200ms prior to the onset of any task stimulus. Trial pupil diameter was measured every 4ms for a fixed 2000ms period starting from the onset of task-relevant stimuli. Specific sequences of trial pupil measurement periods in each task are detailed in the subsequent sections explaining task procedures. Both baseline and trial pupil diameters were measured in pixel values. Trials with more than 50% baseline or trial pupillary data lost due to artifact or blinking were excluded from analyses. Behavioral response time (RT) and accuracy were recorded on each trial in all tasks. Single-trial linear regression/correlation analyses rather than the traditional “baseline subtraction + averaging” approach were employed to analyze pupillary data. As implemented elsewhere (Krishnamurthy, Nassar, Sarode, & Gold, 2016), defining PD evoked by task manipulations after regressing out variances of baseline pupil with multiple regression models is an effective method to control for fluctuations in baseline pupil diameter. This approach yielded independent regression coefficients for task 16 conditions and other task variables, which represented magnitudes of change associated with a unit change in predictors. Critically, the regression coefficients for task manipulations provided the operational definition of pupillary responses as a single time- series response curve for each subject in each task. These response curves were derived after controlling for covariates that could also influence trial pupil diameter. One-sample t-tests were then applied at each time point during the two-second trial periods to examine whether group distributions of each coefficient differed from zero. Temporal plots of significant time points at the group level determined the time window for averaging coefficients at the individual level. Averaging each subject’s regression coefficients in this data-driven time window yielded a single numerical measure of pupillary responses in each task. These averaged measures of pupillary responses were used in age group comparisons. To correct for multiple comparisons in repeated t-tests, cluster-size-based permutation testing with 10,000 iterations was used for significance testing of the overall effects of predictors. Specifically, a cluster size was calculated by summing the number of time points with p < 0.05 across the one-sample t-tests. Then a permutation corrected p-value was calculating by computing the frequency of this cluster size exceeding cluster sizes in a permutation distribution obtained by randomizing the signs of coefficients individually across subjects (Nichols & Holmes, 2001). Phasic Alerting Task In the phasic alerting task, participants completed a visual target localization task with intermittent preceding the auditory tone as illustrated in Figure 1. The task included 17 a practice block of 10 trials with 30% sound trials, and two testing blocks of 100 trials, in which the sound condition were randomly presented on 30% trials in each block without feedback. Each trial began with a fixation point at the center of the screen flanked on each side by two square boxes (3.00° of visual angle sides). The center of each box was positioned 5.30° of visual angle from the fixation cross. On 30% of trials, an auditory tone (white noise, 53 dB, 100ms, Fast Response Weighting C Mode) was presented after a random delay interval between 1,000 and 1,500ms. For the remaining 70% no sound trials, the fixation point stayed on for an additional100ms. Regardless of the sound condition, baseline pupil was measured for the last 200ms of this delay interval prior to onset of the auditory tone. A mean inter-stimulus interval (ISI) of 50ms was included between the end of auditory tones and the onset of targets, which was randomized between 40 and 60ms to produce a mean stimulus onset asynchrony (SOA) of 150ms. The black circle target (diameter of 0.70° of visual angle) was presented at the center of one of the flanking boxes. Trial pupil diameter was measured for 2,000ms starting at the onset of the black circle target. Participants indicated which box contained the target by pressing one of two response keys in corresponding spatial positions. After a response was made, the box frame remained on for an additional 2,000ms followed by 500ms of fixation screen. In multiple regression analyses for the phasic alerting task, only accurate trials were included. The following regression model was applied at each time point for each subject: Trial Pupil Diameter = Intercept + β1Speed + β2Sound + β3ToT + β4SpeedxSound + β5Baseline 18 In the equation, Speed denoted the inverse of behavioral RT; Sound encoded the sound condition with sound being 1 and no sound being 0; ToT stood for time-on-task effect, which was computed with log of trial numbers in each block; SpeedxSound was the interaction effect of speed effects in different sound conditions; and Baseline captured the average of 200ms baseline pupil diameter. Recognition Memory Task In the recognition memory task, participants completed an overt word recognition memory task with immediate recall as illustrated in Figure 2. Two sets of 50 words selected from a picture naming list (Snodgrass & Vanderwart, 1980) were used in counterbalanced order as the learning list and new list. All words had with comparable complexity, familiarity, and frequency counts (p > .1). During the learning phase, each trial began with 2,000ms fixation screen, then a list of 50 words was sequentially presented in a randomized order with each word lasting 2,000ms at the center of the screen. Consistent with a previous study (Võ et al., 2008), all words were capitalized, 3-6 letters long, and subtended a vertical visual angle of 0.69°, resulting in a horizontal visual angle ranging from 1.60° to 3.21°. No response was required during learning; participants were instructed to try their best to remember the list of words and that they would be tested later. The recognition phase followed the learning phase immediately after a short break (about 2min). During recognition, a list of 100 words, comprised of the 50 old words from learning phase and 50 new words, was presented in an identical manner as in the learning phase. Baseline pupil diameter was measured for the last 200ms of fixation, and 19 trial pupil diameter was measured for the 2,000ms word presentation period. At the end of word presentation, a prompt screen was displayed asking whether the word in the trial was “Old” or “New”. The prompt screen stayed on until a response was received. Participants were instructed to press corresponding keys only on the prompt screen but not before, as responses keys were inactivated during word presentation. In multiple regression analyses, the following regression model was applied at each time point in the recognition memory task for each subject: Trial Pupil Diameter = Intercept + β1Speed + β2OldorNew + β3Accuracy + β4ToT + β5SpeedxOldorNew + β6OldorNewxAccuracy + β7SpeedxAccuracy + β8Baseline Speed, ToT, and Baseline stayed the same as in the phasic alerting task equation; OldorNew encoded the whether a word was old or new with old being 1 and new being 0; Accuracy captured whether the judgement was correct with accurate being 1 and inaccurate being 0; and SpeedxOldorNew, OldorNewxAccuracy, and SpeedxAccuracy were a set of interaction effects between the three main effects. Flanker Task In the flanker task, participants completed an arrow version of flanker task (Eriksen & Eriksen, 1974; Mayr, Awh, & Laurey, 2003) as illustrated in Figure 3. The task started with a practice block of 8 trials with 50% congruent trials, followed by one block of 120 trials, in which each condition was presented for half of the trials in a randomized order with no feedback. For each trial, a blank screen was presented for a fixed 1,000ms. Baseline pupil diameter was measured for the last 200ms of the blank screen. The target and flanking arrows then appeared at the center of the screen for 500ms. The arrows were then 20 replaced with a fixation cross for 1,500ms. The target consisted of five horizontal arrows (1.68° x 2.44°) spaced 0.99° apart, consistent with the high conflicting strength condition (Takezawa & Miyatani, 2005). During congruent trials, the five arrows pointed in the same direction; during incongruent trials, the central arrow pointed in the opposite direction from those of flanking arrows. Trial pupil diameter was measured for the 2,000ms combining arrow screen and fixation screen. Participants were instructed to judge the direction of the central arrow by pressing the left or right response keys, as quickly as possible and regardless of the direction of flanking arrows. In multiple regression analyses for the flanker task, only accurate trials were included. The following regression model was applied at each time point for each subject: Trial Pupil Diameter = Intercept + β1Speed + β2Congruency + β3PreviousCongruency + β4ToT + β5SpeedxCongruency + β6Baseline Speed, ToT, and Baseline stayed the same as in the first two tasks; Congruency encoded the congruency condition with incongruent being 1 and congruent being 0; PreviousCongruency encoded the congruency condition on the previous trial with incongruent being 1 and congruent being 0; and SpeedxCongruency was the interaction effect of speed effects in different congruency conditions. 21 Results Phasic Alerting Task Behavioral Mean and standard deviation of accuracies and RTs for the sound and no sound conditions are shown in the first panel of Table 2. All participants achieved ceiling accuracies in the task, and paired-sample Wilcoxon Signed Rank tests revealed no significant difference in accuracies between sound conditions (YC: Z = -1.3716, p = 0.1702; ED: Z = 0.1814, p = 0.8561). Paired samples t-tests revealed significant reductions in RTs by auditory tones in both age groups (YC: t = -7.0362, p < 0.01, d = - 1.4072; ED: t = -9.6734, p < 0.01, d = -1.9347). As shown in Figure4, ED subjects were overall slower than YC on RTs (t = 9.5616, p < 0.01, d = 1.9123). No aging effect was found in mean response accuracy using Mann-Whitney U-test (Z = 0.8950, p = 0.3708). Two-sample t-test showed an increased RT sound effect in ED compared to YC (t = 4.0058, p < 0.01, d = 1.1330), which was still present using sound effects normalized by dividing with no sound condition RTs (t = 2.2403, p = 0.0297, d = 0.6337). Pupillary Responses Figure 5 shows the grand-average PD for each 4ms measurement along the time course of the trial period for each sound condition. The waveforms illustrate increased pupil dilatory sizes in the sound condition compared to the no sound condition, which can be observed in both age groups. Two-sample t-test on average baseline pupil diameter revealed no age group difference (YC baseline = 4066 ± 665 pixels; ED baseline = 4161 22 ± 685 pixels; ± stands for standard deviation; t = 0.4945, p = 0.6232). Similarly, average intercept terms (Figure 6) from regression models across the trial period showed no significant age group difference (t = -0.0290, p = 0.9770), suggesting no intrinsic difference in PD beyond age-related changes in pupillary responses to task variables. Figure 7 illustrates the response curves derived from regression analyses with separate plots for each independent variable included in the linear model. The magnitudes of the unstandardized regression coefficients were plotted across the two-second trial time window. Points marked in red along the response curves indicate B values significantly different from zero. Shaded areas represent standard errors from the mean coefficients. Baseline pupil diameter (Figure 7a) was positively correlated with trial pupil diameter throughout the 2s trial period (permutation test: YC and ED: durations = 2000ms, ps < 0.05). The strength of this correlation decreased across the trial period. Unlike subtracting the same mean baseline from all time points, regression on time-series allowed contributions of baseline to trial pupil diameter to be adjusted across time points. Averages of baseline regression coefficients for each subject were calculated within the significant time points identified at the group level. Two-sample t-test on the average coefficients revealed no significant aging effect on the baseline correlation (t = -0.0385, p = 0.96942). Strong pupillary sound effects (Figure 7b) were identified in both age groups. The auditory tone elicited a significant increase in PD beginning at around 200ms after the target stimulus, and the effect peaked near the time of actual behavioral responses (permutation test: YC: duration = 1664ms, p < 0.01; ED: duration = 1632ms, p < 0.01). Using average coefficients, no aging effect was identified on the pupillary sound effect (t 23 = -1.0028, p = 0.3210). Regression on time series with age group as an independent variable also revealed no significant difference between pupillary sound effects in YC and ED across the 2s time window (ps > 0.05). Response speed (Figure 7c) was positively correlated with trial PD during the early trial period. This relationship peaked at around 550ms (permutation test: YC: duration = 856ms, p < 0.05; ED: duration = 888ms, p < 0.05) and was not significant during the latter second of the trial period. Two-sample t-tests on average coefficients revealed a significant aging effect on the correlation between response speed and PD (t = -3.2929, p < 0.01, d = -0.9314), with ED subjects showing a weakened correlation compared to YC. In addition, the speed effect was influenced by a significant interaction effect (Figure 6d) between response speed and sound condition in YC but not in ED. There was a negative correlation between the interaction term and trial PD in YC, suggesting a reduced correlation with speed in sound trials compared to no sound trials (permutation test: YC: duration = 752ms, p < 0.05). To further investigate this interaction, separate regression models (excluding the sound variable) were performed on the sound and no sound conditions. The observed correlations between response speed and trial PD are shown in Figure 8. The interaction effect in YC was driven by an increase in correlation between speed and PD in the no sound condition. The age group difference in the speed effects and the lack of interaction effect in ED were both due to a lack of increase in correlation between speed and PD in the no sound condition for the ED subjects. Time-on-task effects (Figure 7e) were identified in both age groups. The negative correlations between log of trial number and trial PD suggest gradual decrease in pupil 24 dilatory sizes as time went on during a testing block (permutation test: YC: duration = 1976ms, p < 0.05; ED: duration = 2000ms, p < 0.05). Difference by age group was identified on the time-on-task effects using average coefficients, with ED showing less negative correlations (t = 3.3963, p < 0.01, d = 0.9606) than YC subjects, suggesting an attenuated decrease in PD over time in the ED group. Recognition Memory Task Behavioral Mean and standard deviation of hit rates and false alarm rates are shown in the second panel of Table 2. Non-parametric measures of sensitivity and response bias, A’ and B’’, from Signal Detection Theory (Stanislaw & Todorov, 1999) were also computed and included in the table. Hit rates and correct rejection rates for all subjects are plotted in Figure 9. Paired-sample Wilcoxon Signed Rank tests revealed no difference in hit versus rejection rates in both age groups (YC: Z = 0.4573, p =0.6475; ED: Z = -1.5009, p = 0.1334). Mann-Whitney U-test showed a significant reduction in hit rates (Z = -2.5838, p < 0.01, with 69.6% pairs supporting reduced hit rate) but similar correct rejection rates (Z = -0.1166, p = 0.9072) for ED compared to YC subjects. Correspondingly, Mann- Whitney U-tests revealed reduced sensitivity measured using A’ (Z = -2.2217, p < 0.05, with 68.3% pairs supporting reduced A’) and comparable response bias in terms of B’’ (Z = 1.2712, p = 0.2037). Because participants were instructed to withhold responses until prompt screens, RTs were not analyzed in this task and only included as response speed in the regression models as an independent variable. 25 Pupillary Responses Figure 10 shows the grand-average PD for each 4ms measurement along the time course of the trial period for each word condition during the recognition phase. The waveforms illustrate increased pupil dilatory sizes in old words compared to new words in correct trials in both age groups. Two-sample t-test on average baseline pupil diameter revealed no age group difference (YC baseline = 4296 ± 742 pixels; ED baseline = 4405 ± 766 pixels; ± stands for standard deviation; t = 0.5128, p = 0.6105). Like in the phasic alerting task, no significant age group difference was identified on the average intercept terms (t = 0.4002, p = 0.6908). Figure 11 illustrates the response curves derived from regression analyses during the recognition phase. Plots related to response speed are omitted from this figure. No significant main effect or interaction was found for these speed-related variables, and the exclusion is justified by the recognition memory task design in which participants withheld responses until prompt screens. Baseline and time-on-task effects have identical patterns to those in the phasic alerting task. Both ED and YC subjects showed positive correlations between baseline pupil diameter and trial pupil diameter (Figure 11a) throughout the 2s trial period (permutation test: YC and ED: durations = 2000ms, ps < 0.05; t = 0.2280, p = 0.8206). An aging effect on the time-on-task effects (Figure 11e) was also found in this task (permutation test: YC: duration = 1856ms, p < 0.05; ED: duration = 1604ms, p < 0.05; t = 2.9567, p < 0.01, d = 0.8363). This pattern replicated the attenuated decrease in PD over time for ED compared to YC found in the phasic alerting task. Pupillary old/new effects (Figure 11b) were identified in both age groups. Old words elicited a significant increase in PD compared to new words, which progressively 26 increased over the course of the 2s trial period (permutation test: YC: duration = 1248ms, p < 0.05; ED: duration = 1008ms, p < 0.05). Using average coefficients, no aging effect was identified on the pupillary old/new effect (t = -0.4755, p = 0.6366). Regression on time series with age group as an independent variable also revealed no significant difference between pupillary old/new effects in YC and ED across the 2s time window (ps > 0.05). Accuracy effects (Figure 11c) were absent in both age groups, potentially due to few incorrect trials given the relatively high accuracy rates. However, a significant interaction effect (Figure 11d) of pupillary old/new effect with accuracy condition was found in YC but not in ED. There was a positive correlation between the interaction term and trial PD in YC, suggesting an increased pupillary old/new effect in the correct trials compared to incorrect trials (permutation test: duration = 944ms, p < 0.05). To further investigate this interaction, separate regression models (excluding the accuracy variable) were performed on the correct and incorrect trials. The observed pupillary old/new effects are shown in Figure 12. Significant pupillary old/new effects were identified in the correct trials. Although graphically ED subjects seemed to exhibit a reduced pupillary response, average coefficients revealed no aging effect on the pupillary old/new effect in the correct trials (t = -0.9604, p = 0.3417). Furthermore, no significant pupillary old/new effects were found in the incorrect trials for both age groups. Upon close inspection of the non-significant effects, the fluctuation of pupillary old/new response in YC was in the negative direction, therefore driving the significant interaction effect in terms of greater pupillary old/new effect in correct than incorrect trials. In contrast, the fluctuation of pupillary old/new response in ED was in the positive 27 direction, reaching statistical significance for a few time points around 760ms. Hence, no significant interaction effect was found for the ED group. Flanker Task Behavioral Mean and standard deviation of accuracies and RTs for the congruent and incongruent conditions are shown in the third panel of Table 2. All participants achieved high albeit non-ceiling accuracies in the task. Paired-sample Wilcoxon Signed Rank tests revealed higher accuracies in the congruent condition than the incongruent condition (YC: Z = 3.4786, p < 0.01, with 76.0% pairs showing increased accuracies, 20.0% showing equivalent accuracies; ED: Z = 2.2616, p < 0.05, with 48.0% pairs showing increased accuracies, 28.0% showing equivalent accuracies). Paired samples t-tests revealed significant increases in RTs by the incongruent condition compared to the congruent condition (YC: t = 11.6417, p < 0.01, d = 2.3283; ED: t = 6.6336, p < 0.01, d = 1.3267). As shown in Figure 13, ED subjects were overall slower than YC on RTs (t = 10.9490, p < 0.01, d = 2.1898). No aging effect was found in mean response accuracy using Mann-Whitney U-test (Z = 0.1581, p = 0.8744). Two-sample t-test showed similar congruency effects between age groups (t = -1.8240, p = 0.0735). However, a significant reduction in congruency effect in ED compared to YC was found using congruency effects normalized by dividing with congruent condition RTs (t = -3.8445, p < 0.01, d = - 1.0874). 28 Pupillary Responses Figure 14 shows the grand-average PD for each 4ms measurement along the time course of the trial period for each congruency condition. There was a constriction pattern in trial pupil diameter at around 500ms, which might be driven by an internal effort to focus on the central target arrow. Regardless this initial constriction, the waveforms illustrate increased pupil dilatory sizes in the incongruent condition compared to the congruent condition, which seemed to be observable in both age groups. Two-sample t- test on average baseline pupil diameter revealed no age group difference (YC baseline = 4488 ± 789 pixels; ED baseline = 4274 ± 755 pixels; ± stands for standard deviation; t = - 0.9795, p = 0.3322). No significant age group difference was identified on the average intercept terms (t = -0.8899, p = 0.3780). Figure 15 illustrates the response curves derived from regression analyses with separate plots for each independent variable included in the linear model. Baseline and time-on-task effects have identical patterns to those in the phasic alerting task and recognition memory task. Both ED and YC subjects showed positive correlations between baseline pupil diameter and trial pupil diameter (Figure 15a) throughout the 2s trial period (permutation test: YC and ED: durations = 2000ms, ps < 0.05; t = -0.8059, p = 0.4243). Replicating the previous two tasks, an aging effect on the time-on-task effects (Figure 15e) was found in this task (permutation test: YC: duration = 1736ms, p < 0.05; ED: duration = 1828ms, p < 0.05; t = 2.9706, p < 0.01, d = 0.8402), with ED subjects showing an attenuated decrease in PD over time. Contrary to the raw PD plot (Figure 14), pupillary congruency effect (Figure 15b) was only identified in the YC group but not in ED. Incongruent trials elicited an 29 increased PD compared to congruent trials in YC (permutation test: YC: duration = 1068ms, p < 0.05). The significant time points were broken into two segments. Given the continuous increasing trend of the response curve, the two segments are likely the same effect with increased variability in the middle segment that caused non-significant time points. Because no main effect of congruency was identified for the ED group, the significant time window identified from YC’s congruency effect was used to compute average coefficients. Using these coefficients, no aging effect was identified on the pupillary congruency effect (t = -1.5850, p = 0.1195). Regression on time series with age group as an independent variable also revealed no significant difference between pupillary congruency effects in YC and ED across the 2s time window (ps > 0.05). Considering the large standard error intervals in the YC group, the nonsignificant aging effect might be due to high variability in the pupillary congruency responses. No significant effect of congruency on the previous trial (Figure 15f) was found on the trial PD in both age groups. Response speed (Figure 15c) showed distinct patterns between age groups in this task. In YC subjects, this relationship resembles that in the phasic alerting task, which indicates an early positive correlation between response speed and trial PD. However, this effect was brief and therefore did not survive permutation testing (permutation test: YC: duration = 375ms, p = 0.089). In ED subjects, a late but negative correlation was found between response speed and trial PD (permutation test: ED: duration = 964ms, p < 0.01) and was not significant in the early trial period. Consistent with flipped correlations identified on response curves, two-sample t-tests on average coefficients revealed a highly significant aging effect on the correlation between response speed and PD (t = - 30 4.4956, p < 0.01, d = -1.2716). No significant interaction effect (Figure 15d) was found between response speed and congruency conditions. 31 Discussion This study examined aging effects on task-evoked pupillary responses reflecting LC-NA activities in three tasks from different cognitive domains. The goal of this study was to investigate age-related differences in previously identified PD responses to task manipulations as evidence for functional changes in the phasic responsiveness of the LC- NA system. Given the pervasive impact of LC-NA signaling on cognitive functioning, characterizing alterations in the LC-NA system during task performance could inform underlying mechanisms of cognitive decline during normal aging. To capture the broad modulation of noradrenergic projections, a phasic alerting task, a recognition memory task, and a flanker interference task were used to target different cognitive processes, namely attentional arousal, memory retrieval, and conflict resolution, where reliable PD responses have been established and age-related declines have been reported in the behavioral literature. Through administering tasks with constant luminance level and central fixation, this study controlled for potential confounding factors in assessing pupillary responses across age groups. The use of a regression approach to pupillary data provided further advantages of isolating PD responses to task manipulations and extending pupillary responses beyond main effects to correlations with behavioral performance and their interactions. In summary of aging effects on the profile of pupillary responses found in this study, comparable PD responses to task manipulations were identified in YC and ED subjects during the phasic alerting task and recognition memory task. Both age groups showed an increased PD in response to auditory warning cues preceding a target localization task, reflecting an elevated arousal state elicited by the cues. Similarly, both 32 age groups had increased PD responses to old items compared to new items during retrieval in a recognition memory task, suggesting altered LC-NA signaling during retrieval of stored words versus rejecting new words. Despite preserved magnitudes of pupillary responses to sound tones in the phasic alerting task and to old words in the recognition memory task, ED subjects showed qualitatively different patterns of PD responses from those of the YC group in relation to behavioral performance, potentially suggesting reduced effectiveness of the LC-NA system to benefit overt responses. Lastly, a main effect of pupillary congruency response as greater PD in incongruent trials than congruent trials in the flanker interference task was only identified in YC subjects, but not in the ED group. Reduced phasic LC-NA responses during conflict resolution might be caused by specific mechanistic engagement of noradrenergic projections in cognitive control processes distinct from LC-NA contributions in attentional arousal and memory retrieval functions. Overall, findings in this study demonstrated some preserved patterns of LC-NA signaling in the aging population while also pointing out other age-related changes in the LC-NA system within the context of cognitive tasks. The use of three tasks in different cognitive domains accommodated the heterogeneous nature of LC-NA signaling as indexed by PD responses in different cognitive processes. Finally, aging effects on altered correlations between trial PD and behavioral performance identified in this study highlighted the benefits of using partial pupil dilatory response curves to characterize PD responses during cognitive task performance. In the phasic alerting task, YC subjects showed a robust increase in PD in the sound condition compared to the no sound condition. This effect was captured by coefficients of the sound variable in the regression models, and the significant effect 33 lasted for most of the 2s trial period (Figure 7b). The increased PD suggested a phasic burst of LC-NA signaling triggered by the auditory warning cues, reflecting an increased arousal state as the result of the cues. In accordance with the view that an elevated arousal level leads to faster sensory processing of target stimuli and behavioral actions (Petersen & Posner, 2012), a significant reduction in RTs was identified in the sound condition in YC subjects. The same pattern of pupillary responses to sound tones and speeded behavioral responses have been shown in several previous studies (Gabay et al., 2011; Geva et al., 2013; Tona et al., 2016). In line with a LC-NA-mediated neural implementation of arousal, the pupillary sound effect has been consistently interpreted as a psychophysiological marker of a phasic alerting effect within the attentional network. In addition to the main effect of sound, a significant positive correlation between response speed and trial PD during the first second of the trial period was also identified in YC subjects (Figure 7c). It needs to be clarified that the correlated trial PD is different from the PD response elicited by sound tones because both sound and response speed were simultaneously entered in the regression equation. Hence, the trial PD here referred to dilation in pupil sizes during the trial time window beyond dilations due to sound; such trial PD likely captured normal fluctuations in LC-NA signaling and increased LC-NA signaling in response to the presentation of target stimuli and preparation of behavioral responses. A positive correlation between response speed and trial PD indicated higher pupil dilatory sizes in trials with faster behavioral responses of deciding the locations of target stimuli. This correlation is consistent with the LC-NA model of arousal (Aston- Jones & Cohen, 2005; Petersen & Posner, 2012), where increased LC-NA signaling is thought to enable speeded behavioral responses. In support of this account, it has been 34 previously noted that LC-NA activation might be more temporarily related to behavioral responses rather than the presentation of task stimulus itself (Rajkowski, Majczynski, Clayton, & Aston-Jones, 2004). Then, the positive correlation between trial PD and response speed in this study might be more specifically attributed to the process of behavioral response preparation, which was mediated by a phasic burst of LC-NA signaling. Indeed, a monkey study has suggested that task-evoked LC-NA activity is relevant for energizing a response behavior by enhancing sensory-motor processes (Varazzani, San-Galli, Gilardeau, & Bouret, 2015). The main effect of sound in YC subjects was further modulated by an interaction with sound conditions (Figure 7d). As shown in separate modeling of the correlations in each sound condition (Figure 8 top panel), this interaction effect was driven by a higher association strength in the no sound condition compared to the sound condition. In other words, the same magnitude of trial PD was correlated with a greater increase in behavioral response speed. One possible explanation is that in the no sound condition, response speed was not modulated by auditory warning cues; therefore, intrinsic LC-NA signaling was the main determinant of the behavioral response preparation process. Intrinsic is used here to indicate PD responses to task stimuli presentations as opposed to manipulated sound conditions. Then, the association between trial PD and response speed in the no sound condition could be viewed as a default effectiveness of the LC-NA system in mediating speeded behavioral responses. Contrary to our prior expectation, ED subjects showed a comparable pupillary sound effect as in the YC group (Figure 7b). Age group comparison using average coefficients of the sound variable found no significant aging effect on increased PD in 35 response to auditory warning cues. In addition, t-tests on age group differences across individual time points during the 2s trial period also found no time point with a significance level exceeding the α level at 0.05. The time courses of the significant pupillary sound effect were almost identical between the two age groups. Mirroring the behavioral pattern in YC subjects, ED subjects had significantly reduced RTs in the sound condition compared to the no sound condition. These results showed that ED subjects had a preserved phasic burst of LC-NA signaling in response to the auditory warning cues in the phasic alerting task, suggesting a largely intact arousal network. Therefore, the attentional alerting effect elicited in the sound condition was mediated by noradrenergic projections that were functionally intact with normal aging. Response speed and trial PD were also positively correlated with each other for ED subjects in the phasic alerting task (Figure 7c). Significant coefficients of the speed variable were observed in the early time window, peaking at around 500ms, which was comparable with the time course seen in YC subjects. However, the correlation between response speed and trial PD was significantly weaker in the ED group compared to the YC group. This suggested that even though on average pupil dilatory sizes were higher in trials with faster response speed, the association between LC-NA signaling and response behaviors was weakened during normal aging. Furthermore, no significant interaction of the response speed correlation with sound conditions was found in the ED group. Separate modeling in each sound condition identified that both the reduced correlation between response speed and trial PD and the absence of significant interaction effect were driven by an attenuated association between trial PD and response speed in the no sound condition for ED subjects, while the associations were similar between age groups 36 in the sound condition (Figure 8 bottom panel). In the no sound condition, the behavioral response preparation process was not mediated by auditory warning cues. Therefore, the correlation provided a measurement of how tightly linked together are LC-NA signaling and speeded responses during cognitive functioning in a simple target localization task. Given the LC-NA model of arousal, the identified aging effect on the association between LC-NA signaling and response behaviors highlighted a subtle age-related change in the LC-NA system that was complementary to the preserved pupillary sound effect described above. On the one hand, it was found that ED subjects had an intact phasic attentional alerting response to auditory warning cues, which increased arousal levels and translated into faster responses. On the other hand, the results showed that when no external signal was available, the LC-NA system in ED subjects was less effective compared to YC subjects in utilizing intrinsic LC-NA signaling to influence behavioral response speed. This pattern of decoupling between a neuromodulation system and behavioral performance might be the result of increased neural noise in downstream sensory-motor processes. Hence, the arousal network in the ED group could still be largely intact with respect to producing a burst of noradrenergic projections in response to salient auditory warning cues; however, in the absence of a bulk of NA-signaling, the influence of intrinsic fluctuations of the LC-NA system on the response preparation process was weakened during normal aging. This pattern was corroborated by aging-effect results from the recognition memory task in this study. The theme of the LC-NA system with an intact task-evoked main effect with aging yet reduced efficacy in influencing behavioral performance was echoed in the distinct cognitive process of memory retrieval. 37 In the recognition memory task, YC subjects showed the typical pupillary old/new effect, where PD was greater for old words compared to new words during the recognition phase. This effect replicated the old/new effects reported in other studies (Brocher & Graf, 2017; Goldinger & Papesh, 2012; Heaver & Hutton, 2011; Kafkas & Montaldi, 2015; Otero et al., 2011; Võ et al., 2008). The derived response curve added to the literature by charactering the time course of the pupillary response as emerging later during the 2s trial period and continuously increasing until the end of the trial. This pattern was explained by the memory-strength account, which argued that the increased PD responses reflect greater memory strength associated with the old items compared to new items (Otero et al., 2011; Papesh et al., 2012; van Rijn, Dalenberg, Borst, & Sprenger, 2012). It should be noted that the memory-strength account is a purely cognitive theory, which does not interpret the increased PD as a proxy measure for LC- NA signaling, and the LC-NA system was not mentioned as a neural mechanism involved in the memory retrieval process. However, there is an extensive literature on NA-based neuromodulation in memory functioning. Previous studies have reported that context cues might facilitate memory retrieval through increased arousal mediated by LC-NA activities (Sara, 2000; Sara & Devauges, 1988; Sara & Hars, 2006), and NA-signaling has been shown to facilitate signal-to-noise ratios during sensory processing (Bouret & Sara, 2002; Lecas, 2004). Therefore, it is likely that retrieving the memory of old items in the recognition memory task elicited a phasic increase in LC-NA signaling when the bottom-up visual stimuli of words interacted with encoded memory traces of the old words. This form of interaction has been previously termed ecphory (Tulving, 1982). It is possible that the wide noradrenergic projections then enhanced synchronous spiking of 38 neurons involved in the distributed representations of old-word memories by mediating the signal-to-noise ratio in sensory areas. Consequently, facilitation of the memory retrieval process resulted in a behavioral decision of judging the presented word as an old word. Consistent with this neurobiological version of the memory-strength account, the pupillary old/new effect in YC subjects was modulated by an interaction with the accuracy variable in the regression models. As shown in Figure 11d, positive coefficients for the old/new by accuracy interaction variable indicated a greater pupillary old/new effect in the accurate trials compared to the inaccurate trials. On the one hand, new words incorrectly identified as old words might reflect an erroneous activation of stored memory, which would also elicit increased LC-NA signaling to some extent. On the other hand, old words incorrectly identified as new words could indicate weak memory traces. Since LC-NA signaling was likely a response to the activation of memory traces, when a memory trace was weak potentially due to noisy encoding, interaction of such memory trace with visual stimuli during retrieval should elicit a weaker LC-NA signaling. Combing the two cases within inaccurate trials, one should expect a reduced pupillary old/new effect. Indeed, separate modeling of the pupillary response in inaccurate trials showed that the pupillary old/new effect was non-significant throughout the 2s trial period with a trend deflecting in the opposite direction as in the accurate trials (Figure 12 top panel). This result suggested that the pupillary old/new effect was restricted to successful retrieval of encoded memory traces, as predicted by the memory-strength account. 39 As expected with memory decline during normal aging, ED subjects performed significantly worse than YC subjects in the recognition memory task. Using signal- detection-theory measures, ED subjects showed significantly reduced sensitivity to old versus new words during retrieval, while response biases were comparable between age groups. However, contrary to the behavioral deficit, a preserved pupillary old/new effect was identified in the ED group. No age group difference was found neither using average coefficients nor t-tests at individual time points along the trial period (Figure 11b). Although the duration of the significant old/new effect was shorter in ED subjects compared to the YC group, the overall shape of response curve of the pupillary response to word condition was similar in both age groups. This finding suggested that old words still elicited an increased LC-NA signaling compared to new words, reflecting stronger memory strength for encoded old word memories. Despite worse behavioral memory performance compared to YC subjects, ED subjects were able to retrieve old word memories better than chance in this task. Therefore, the preserved pupillary old/new effect was still compatible with the memory-strength account. On trials when the retrieval cues successfully interacted with encoded memory traces, LC-NA signaling increased to a greater extent compared to new word trials. Because of the use of response curves derived from regression models, the coefficients for the old/new variable captured the difference in LC-NA signaling between old and new words independent of response accuracy. What the result did reveal was that the reduced memory performance with aging was not due to a general impairment in the functional phasic responses of the LC- NA system, since old words elicited the same magnitude of increase in LC-NA signaling as in YC subjects. Given the identified intact LC-NA system for producing a pupillary 40 old/new effect during normal aging, cognitive decline in memory performance might be the result of age-related changes in cognitive processes other than the NA-mediated arousal network. One possible cognitive process affected by aging might be the decision-making stage of behavioral response. Investigation of the interaction effect of the pupillary old/new effect with response accuracy showed that there was a lack of significant interaction effect in ED subjects. Separate modeling in inaccurate trials identified a response curve deflecting in the same direction as in the accurate trials (Figure 12 bottom panel), and the effect reached significance level briefly at around 750ms albeit not surviving a permutation testing correction. This result suggested that compared to YC subjects, the pupillary old/new effect was less discriminative between correct versus incorrect trials in the ED group. In other words, the increased LC-NA signaling was no longer restricted to correctly recognized old words. This pattern could indicate a noisy decision-making process that would incorrectly classify memory retrieval instances that elicited increased LC-NA signaling as new words, despite the successful interaction between retrieval cues and encoded memory traces. The Synergistic Ecphory Model of Retrieval provided a relevant insight regarding the distinction between the recollection of a memory and the overt report of the recollective experience of the memory (Tulving, 1982). In Tulving’s model, ecphoric interaction was only one of two processes involved in successful memory retrieval, and the other process was termed conversion. Conversion captured the stream of cognitive processes recruited in translating the conjunction of encoded memory and retrieval information into an overt behavioral response. In the experimental paradigm of the recognition memory task, conversion included the 41 decision-making process that produced the button-press response behavior that indicated whether a presented word was old or new. Hence, even though old words elicited an increased PD compared to new words, there could be a reduced impact of the noradrenergic projections on downstream processes within the sensory-motor areas involved in decision-making. Specific NA modulation on sensory information has been outlined in the process of memory encoding (Clewett, Huang, Velasco, Lee, & Mather, 2018; Hoffing & Seitz, 2015), but given the sensory-reactivation hypothesis (Wheeler, Petersen, & Buckner, 2000), the same mechanisms could modulate decision making during memory retrieval. Furthermore, phasic arousal has been shown to dynamically suppress decision bias, allowing a more evidence-driven response in choice tasks (de Gee et al., 2017). Therefore, age-related changes in the effectiveness of the LC-NA system in modulating cognitive processes involved in conversion might underlie the reduced recognition accuracy and sensitivity in the ED group. Replicating the preserved-impaired doublet pattern reported in the phasic alerting task, results from the recognition memory task described another cognitive domain where the LC-NA system during normal aging was intact for a main effect of pupillary response to task conditions. Yet, a subtle age- related change was observed in the relationship between PD responses and overt behavioral performance. In the flanker interference task, YC subjects showed the expected increased PD response to decision conflicts between the central arrow and flanking arrows. As shown in Figure 15b, there was an increased PD in the incongruent trials compared to congruent trials, and the effect had an increasing trend during the 2s trial period. The pupillary response suggested increased arousal in response to conflict detection in the visual 42 stimuli, and a phasic burst of LC-NA signaling indicated effortful cognitive processing and engagement of cognitive control as reported in previous studies (Scharinger et al., 2015; van Bochove et al., 2013; van Steenbergen & Band, 2013). According to the Adaptation By Binding model of conflict processing, the phasic arousal response was elicited after conflict detection, and it indicated an increased demand for cognitive control (Abrahamse et al., 2016). Such arousal response mediated by noradrenergic projections was likely recruited by the frontal conflict processing network. The Expected Value of Control (EVC) theory suggested that dorsal anterior cingulate cortex (dACC) monitors for the presence of conflict during cognitive processing and specifies the control signal for downstream efferent processes (Shenhav, Botvinick, & Cohen, 2013). Given the broad effect of enhancing signal-to-noise ratio by LC-NA signaling (Bouret & Sara, 2002), the LC-NA system might be one of the downstream efferent networks to implement cognitive control in sensory-motor areas. Converging evidence of the role of LC-NA signaling in conflict processing comes from studies showing the influence of pharmacological manipulations of NA concentration on maintenance of attentional focus among distractors (Tiplady et al., 2005), as well as from pupillometry studies showing the role of LC-NA signaling in biased decision-making (de Gee, Knapen, & Donner, 2014). Apart from the main effect of pupillary congruency response, YC subjects showed a positive correlation between trial PD and response speed in the flanker interference task (Figure 15c). Although the significant effect was non-significant after permutation correction for multiple comparisons, the time course as well as the shape of response curve mirrored the pattern observed in the phasic alerting task, in which a greater intrinsic LC-NA signaling was associated with faster behavioral responses. This pattern 43 was largely consistent with the LC-NA model of arousal, where a phasic burst of LC-NA signaling facilitated the behavioral response preparation process and resulted in faster response speed. Hence, despite the fact that the flanker task was mainly dependent on a conflict resolution process, response behaviors of YC subjects were modulated by the overall arousal level, potentially suggesting the ability of YC subjects to adjust to the task demand and to treat the task as a simple target discrimination task with responsive LC- NA signaling to incongruent task stimuli. In contrast, ED subjects did not show a significant pupillary response to decision conflicts in the flanker interference task. As shown in Figure 15b, there was no difference in PD responses in the incongruent trials compared to the congruent trials. At first glance, this result seemed to suggest that ED subjects were not responding in terms of LC-NA signaling to the presence of conflicts. One possible explanation for the lack of main effect of congruency in the pupillary response might be due to reduced connectivity between dACC and LC regions such that conflict signals did not elicit a similar response in the LC-NA system as in YC subjects. However, behavioral performance suggested that ED subjects showed comparable congruency effects as YC subjects; in fact, when normed congruency effects by dividing by the congruent condition RTs were used, ED subjects showed a significantly lower congruency effect compared to YC subjects, indicating a better inhibitory functioning (although ceiling effect must be considered). Hence, it was unlikely that ED subjects failed to implement cognitive control in sensory-motor areas via noradrenergic projections, since such failure would have been reflected as a greater behavioral congruency effect. A more likely explanation is that ED subjects employed a strategy shift in the flanker interference task and intentionally suppressed speeded 44 responses in order to avoid errors. In light of this view, the LC-NA signaling was no longer responsive to incongruent arrows during task stimuli presentation in the ED subjects as characterized in previous literature and in YC subjects in this study. The LC- NA system served as a continuous modulation of arousal level and behavioral responses rather than an arousal response to conflict signals monitored and specified by dACC. This result might be related to the literature on inhibitory functioning in aging population where comparable performance in trials requiring cognitive control was achieved by exerting more overall control compared to YC subjects (Hsieh, Wu, & Tang, 2016; Rey- Mermet & Gade, 2017). These compensatory processes and strategy changes might also contribute to altered functional connectivity patterns and speed-accuracy tradeoff relationships identified during normal aging (Forstmann et al., 2011; Langner et al., 2015). The speculation of a strategy-shift in ED subjects was supported by results of the response speed variable in the regression models. Distinct from the response curves in YC subjects as well as from the pattern from the same ED subjects in the phasic alerting task, trial PD was negatively associated with response speed in the flanker interference task. The time course of the significant negative correlation was also shifted to the second half of the trial period (Figure 15c). The negative correlation coefficients suggested that greater PD responses was associated with slower response behaviors. This relationship was counterintuitive to the LC-NA model of arousal, where increased noradrenergic projections would mediate faster responses. However, the results could be reconciled by considering a strategy-shift in the ED subjects. An elevated arousal state, rather than increasing response speed, slowed down the decision-making process regardless of the 45 congruency condition to allow for a more deliberated response behavior. If such general strategy was employed throughout the flanker task, then response speed alone captured the majority of variance in PD responses and competed against a pupillary congruency effect both during task performance and in the regression models of unique contributions of independent variables. Indeed, in follow-up analyses when response speed was removed from the regression equation, a significant pupillary congruency effect emerged in the ED subjects (permutation test: ED: duration = 928ms, p < 0.01). Together, results of aging effects in the flanker task suggested that LC-NA signaling might be more influenced by a top-down modulation of arousal level in the ED subjects, potentially as the result of a strategy-shift employed to cope with the demand of inhibitory control in a conflict processing task paradigm. Hence, the flanker interference task captured a distinct cognitive process from attentional alerting and memory retrieval, where the phasic response of the LC-NA system was similar between age groups after controlling for shared variances with other task features. One possible difference between conflict resolution and the other two cognitive processes is the nature of a secondary role of LC-NA signaling in the task- dependent conflict processing, which has been more strongly associated with the frontal control networks. While the phasic alerting response to warning cues and ecphory interactions in memory retrieval might trigger phasic LC-NA signaling obligatorily, a pupillary response within the cognitive control domain has been thought to reflect a more general increase in demands of control and processing difficulty (van der Wel & van Steenbergen, 2018); even in the speculated dACC-LC network for exerting cognitive control in posterior cortical regions, noradrenergic projections were more malleable to 46 strategy choices and overall arousal level, since ED subjects might adopt the strategy of exerting indiscriminative control on all responses regardless of visual stimuli as suggested in a Go/No go paradigm (Hsieh et al., 2016). Nevertheless, the pupillary congruency effect could be identified when response speed was removed from the regression equation, suggesting a crude preservation of the pupillary congruency effect in response to the presence of decision conflicts during cognitive task performance. Overall, pupillary responses to task manipulations in the three tasks included in this study showed a functionally intact LC-NA system in the aging population, which contrasts with neurobiological evidence of aging effects on LC-NA signaling. Existing studies have reported age-related degradation of LC nuclei during normal aging, from the early work showing reduced LC cell counts (Mann, 1983; Spokes, 1979; Vijayashankar & Brody, 1979) to more recent studies showing increased LC neurofibrillary pathology in the aging population (Grudzien et al., 2007; Mather & Harley, 2016). Furthermore, the development of MR neuroimaging of neuromelanin, a metabolic byproduct of NA, has inspired investigations of LC-integrity with neuromelanin signal intensity as an in vivo biomarker of LC structure in humans. In the first MRI study of neuromelanin, the LC integrity estimated from signal contrasts to adjacent pontine tegmentum region was found to be significantly decreased in elderly subjects (Shibata et al., 2006). More recently, neuromelanin signals have been shown to correlate linearly with cognitive reserve measures including education and verbal intelligence (Clewett et al., 2016). This finding is especially important for the assessment of the LC-NA system during normal aging, as the linear relationships were consistent with the predictions of the noradrenergic hypothesis of cognitive reserve (Robertson, 2013). Finally, the structural LC integrity 47 measured from neuromelanin imaging has also been shown to be related to memory functioning, adding to the functional relevance of the imaging assessment of the LC-NA system (Hämmerer et al., 2018). Although neuromelanin might be indicative of functional integrity of the LC-NA system as suggested in this last study, the majority of evidence showing age-related decline in LC integrity has been heavily dependent on neurobiological measurements of either LC cell counts or NA-related biomolecular signals. Pupillometry, in comparison, is a much coarser indirect proxy measure of LC- NA activities. However, pupillary responses have the advantage of being originated in cognitive science; therefore, pupillary responses to task manipulations often have direct functional interpretations in relation to the involved cognitive processes. While there might be an overall degradation of LC integrity with normal aging, the functional responsiveness of the LC-NA system could still be largely intact until a substantial proportion of cells have been damaged either by age-related changes or early Alzheimer’s disease pathologies (Grudzien et al., 2007). However, an important complication when using task-evoked pupillary responses as an index for LC-NA signaling is the heterogeneous nature of pupillary responses. Depending on the specific task paradigms, elderly participants have been shown to produce reduced, comparable, or increased PD responses to task manipulations. In relation to the cognitive domains focused in this study, Kim et al. found a comparable dilatory response between age groups in a saccadic target detection task, and the authors interpreted the pattern as a preserved arousal response with aging (Kim et al., 2000). With respect to memory functioning, Hammerer et al. identified similar pupillary old/new effects between age groups, suggesting a preserved phasic increase in LC-NA signaling 48 in response to retrieving encoded old memories (Hämmerer et al., 2017). Opposite patterns were found with cognitive effort manipulations: reduced pupillary responses to increased memory load were found in one study (Van Gerven et al., 2004), while increased pupillary responses to increased sentence complexity in speech comprehension were found in another (Piquado et al., 2010). Hence, no consistent pattern emerged regarding the actual functional integrity of the LC-NA system as indexed by pupillometry measures. A notable caveat in interpreting pupillometry measures is a potentially reduced range of PD responses as results of weakened dilatory muscles with aging, especially when there are changes in luminance levels (Gilzenrat et al., 2010; Moloney et al., 2006; Peysakhovich et al., 2017; Van Gerven et al., 2004). The current study controlled for display luminance by using isoluminant task stimuli and background colors, but it is still important to characterize any difference in overall PD sizes between age groups. The grand-average PD waveforms in all three tasks (Figure 5, Figure 10, Figure 14) seem to show a trend of attenuated PD magnitudes in the ED group compared to YC. However, using coefficients for the intercept terms in regression models averaged across the 2s trial period, no significant age group difference was identified in each task. This pattern suggested that when all independent variables in the regression models including baseline pupil diameter were set to zero, both age groups had similar trial pupil diameter measures (Figure 6), indicating no difference in overall PD sizes beyond the effects captured by each of the independent variables for task features. The current comparison between age groups on trial PD derived from the regression models is distinct from the main effect of a between-subject age group variable used in an ANOVA approach to pupillary data 49 (Ayasse, Lash, & Wingfield, 2017; Carminati & Knoeferle, 2016; Ziaei et al., 2015). In ANOVA, a significant main effect of age group on pupil dilatory sizes averaged across task conditions might be driven by aging effects on pupillary responses to several task features as illustrated in the grand-average PD waveforms in this study. The multiple regression approach employed here provided a statistical alternative to characterize task- independent trial pupil diameter during the presentation of task stimuli, which would be an important preliminary analysis in order to interpret age-group comparisons of task- evoked pupillary responses. Another issue in investigating aging effects on pupillometry measures is that baseline pupil diameter was often neglected in previous studies. A potential difference in baseline pupil diameter between age groups would be critical to report if the baseline mean was subtracted to compute pupil dilatory sizes as recommended in the traditional analytic approach to pupillary data (Beatty & Lucero-Wagoner, 2000). In addition, differences in the baseline pupil diameter might reflect different levels of tonic arousal between age groups (Reimer et al., 2014; Unsworth & Robison, 2016; Van Den Brink et al., 2016). In the current study, t-tests on average baseline pupil diameter during the 200ms time window preceding task stimuli in all three tasks showed no significant difference between age groups. The consistent lack of difference across three tasks suggested that tonic arousal as measured using baseline pupil diameter did not differ significantly between YC and ED subjects in this study. This finding differed from reduced baseline pupil diameter in elderly participants as reported in other studies (Kim et al., 2000; Piquado et al., 2010; Van Gerven et al., 2004). The unusual comparable 50 baseline pupil diameter between age groups might indicate a limited generalizability of findings from ED subjects recruited in this study. Across the three tasks included in the current study, the ED group generally maintained high cognitive functioning in terms of behavioral performance. While ED subjects showed general slowing in RTs in both the phasic alerting task and the flanker interference task, preserved behavioral alerting effects and congruency effects were identified in comparison to YC subjects. This pattern contrasted with results from previous studies of reduced alerting effects of warning cues on RTs (Festa et al., 2004; Tales et al., 2002) and increased congruency effects of response conflicts in a Simon task (Castel et al., 2007). Although a recent meta-analysis only reported sparse findings of inhibition deficits with aging assessed across a range of task paradigms (Rey-Mermet & Gade, 2017), a smaller normed congruency influence on RTs than YC subjects in this study still alluded to superior conflict processing above the level expected with typical aging. Indeed, the demographics of subjects showed on average high years of education in the ED group (Table 1), and only neuropsychological testing dependent on working memory found significant age-group differences (the sequence condition in the Digit Span Task, the switching instruction in the Verbal Fluency task, and the alternating condition in Trails B). The Trails tasks might have captured general slowing in behavioral responses as well, as ED subjects were also slower in Trails A. Furthermore, ED subjects showed high verbal IQs, high cognitive reserve estimates, and near-upper- quartile scores on RBANS. These neuropsychological testing results strongly suggested that the ED subjects participated in this study might be unrepresentative of the general aging population, as they exhibited superior cognitive functioning in multiple assessment 51 scales. Finally, coefficients for the time-on-task variable in the regression models used to analyze pupillary data in all three cognitive tasks were found to be significantly less negative in ED subjects compared to the YC group. This finding suggested that ED subjects were not only highly functional, but also highly motivated, because a decrease in trial PD over time might reflect a habituation to the task stimuli and a reduction in overall cognitive engagement during task performance. Hence, intact LC-NA signaling with aging found in this study might be in part driven by the exceptional cognitive abilities in the ED group, who might be better characterized as “superagers” with preserved neuroanatomy including the LC-NA network (Sun et al., 2016). Future studies are needed to expand the characterization of pupillary response curves derived from regression analyses reported here to elderly participants with a wider range of cognitive functioning. Considering the promising technique of assessing in-vivo LC integrity using MR neuromelanin signals, a combination of LC-neuromelanin imaging and pupillometry would provide a more comprehensive understanding of age-related changes in the functional role of the LC-NA system in cognitive task performance. 52 Reference Abrahamse, E., Braem, S., Notebaert, W., & Verguts, T. (2016). Grounding cognitive control in associative learning. Psychological Bulletin, 142(May), 693–728. Allard, E. S., Wadlinger, H. A., & Isaacowitz, D. M. (2010). Positive gaze preferences in older adults: assessing the role of cognitive effort with pupil dilation. Aging, Neuropsychology, and Cognition, 17(3), 296–311. Arnsten, A. F. T. (1993). Catecholamine mechanisms in age-related cognitive decline. Neurobiology of Aging, 14(6), 639–641. Arnsten, A. F. T., & Goldman-Rakic, P. S. (1985). Alpha 2-adrenergic mechanisms in prefrontal cortex associated with cognitive decline in aged nonhuman primates. Science, 230(4731), 1273–1276. Arnsten, A. F. T., & Rubia, K. (2012). Neurobiological circuits regulating attention, cognitive control, motivation, and emotion: disruptions in neurodevelopmental psychiatric disorders. Journal of the American Academy of Child & Adolescent Psychiatry, 51(4), 356–367. Aston-Jones, G., & Cohen, J. D. (2005). AN INTEGRATIVE THEORY OF LOCUS COERULEUS-NOREPINEPHRINE FUNCTION: Adaptive Gain and Optimal Performance. Annual Review of Neuroscience, 28(1), 403–450. Ayasse, N. D., Lash, A., & Wingfield, A. (2017). Effort not speed characterizes comprehension of spoken sentences by older adults with mild hearing impairment. Frontiers in Aging Neuroscience, 8, 329. Baddeley, A. D., Baddeley, H. A., Bucks, R. S., & Wilcock, G. K. (2001). Attentional control in Alzheimer’s disease. Brain : A Journal of Neurology, 124(Pt 8), 1492– 1508. Beatty, J., & Lucero-Wagoner, B. (2000). The pupillary system. Handbook of Psychophysiology (2nd Ed.). Bergamin, O., & Kardon, R. H. (2002). Greater pupillary escape differentiates central from peripheral visual field loss. Ophthalmology, 109(4), 771–780. Bernstein, I. H., Chu, P. K., Briggs, P., & Schurman, D. L. (1973). Stimulus intensity and foreperiod effects in intersensory facilitation. The Quarterly Journal of Experimental Psychology, 25(2), 171–181. Bernstein, I. H., Rose, R., & Ashe, V. (1970). Preparatory state effects in intersensory facilitation. Psychonomic Science, 19(2), 113–114. Berridge, C. W., & Waterhouse, B. D. (2003). The locus coeruleus-noradrenergic system: Modulation of behavioral state and state-dependent cognitive processes. Brain Research Reviews, 42(1), 33–84. Botvinick, M. M., Cohen, J. D., & Carter, C. S. (2004). Conflict monitoring and anterior cingulate cortex: An update. Trends in Cognitive Sciences, 8(12), 539–546. Bouret, S., & Sara, S. J. (2002). Locus coeruleus activation modulates firing rate and temporal organization of odour-induced single-cell responses in rat piriform cortex. European Journal of Neuroscience, 16(12), 2371–2382. Bradshaw, J. L. (1969). Background light intensity and pupillary response in a reaction time task. Psychonomic Science, 14(6), 271-. Braver, T. S., & Barch, D. M. (2002). A theory of cognitive control, aging cognition, and 53 neuromodulation. Neuroscience & Biobehavioral Reviews, 26(7), 809–817. Breen, L. A., Burde, R. M., & Loewy, A. D. (1983). Brainstem connections to the Edinger-Westphal nucleus of the cat: a retrograde tracer study. Brain Research, 261(2), 303–306. Brocher, A., & Graf, T. (2017). Decision-related factors in pupil old/new effects: Attention, response execution, and false memory. Neuropsychologia, 102, 124–134. Buckner, R. L. (2004). Memory and executive function in aging and AD: multiple factors that cause decline and reserve factors that compensate. Neuron, 44(1), 195–208. Campbell, K. L., Grady, C. L., Ng, C., & Hasher, L. (2012). Age differences in the frontoparietal cognitive control network: Implications for distractibility. Neuropsychologia, 50(9), 2212–2223. Carminati, M. N., & Knoeferle, P. (2016). Priming Younger and Older Adults’ Sentence Comprehension: Insights from Dynamic Emotional Facial Expressions and Pupil Size Measures. The Open Psychology Journal, 9(1). Castel, A. D., Balota, D. A., Hutchison, K. A., Logan, J. M., & Yap, M. J. (2007). Spatial attention and response control in healthy younger and older adults and individuals with Alzheimer’s disease: Evidence for disproportionate selection impairments in the simon task. Neuropsychology, 21(2), 170–182. Clewett, D. V, Huang, R., Velasco, R., Lee, T.-H., & Mather, M. (2018). Locus coeruleus activity strengthens prioritized memories under arousal. Journal of Neuroscience, 38(6), 1558–1574. Clewett, D. V, Lee, T.-H., Greening, S., Ponzio, A., Margalit, E., & Mather, M. (2016). Neuromelanin marks the spot: identifying a locus coeruleus biomarker of cognitive reserve in healthy aging. Neurobiology of Aging, 37, 117–126. Coull, J. T., Büchel, C., Friston, K. J., & Frith, C. D. (1999). Noradrenergically Mediated Plasticity in a Human Attentional Neuronal Network. NeuroImage, 10(6), 705–715. de Gee, J. W., Colizoli, O., Kloosterman, N. A., Knapen, T., Nieuwenhuis, S., & Donner, T. H. (2017). Dynamic modulation of decision biases by brainstem arousal systems. Elife, 6. de Gee, J. W., Knapen, T., & Donner, T. H. (2014). Decision-related pupil dilation reflects upcoming choice and individual bias. Proceedings of the National Academy of Sciences, 111(5), E618–E625. Devauges, V., & Sara, S. J. (1991). Memory retrieval enhancement by locus coeruleus stimulation: evidence for mediation by ??-receptors. Behavioural Brain Research, 43(1), 93–97. Dragan, M. C., Leonard, T. K., Lozano, A. M., McAndrews, M. P., Ng, K., Ryan, J. D., … Hoffman, K. L. (2017). Pupillary responses and memory-guided visual search reveal age-related and Alzheimer’s-related memory decline. Behavioural Brain Research, 322, 351–361. Eckstein, M. K., Guerra-Carrillo, B., Singley, A. T. M., & Bunge, S. A. (2017). Beyond eye gaze: What else can eyetracking reveal about cognition and cognitive development? Developmental Cognitive Neuroscience, 25, 69–91. Elam, M., Thorén, P., & Svensson, T. H. (1986). Locus coeruleus neurons and sympathetic nerves: Activation by visceral afferents. Brain Research, 375(1), 117– 125. Eriksen, B. A., & Eriksen, C. W. (1974). Effects of noise letters upon the identification of 54 a target letter in a nonsearch task. Attention, Perception, & Psychophysics, 16(1), 143–149. Festa-Martino, E., Ott, B. R., & Heindel, W. C. (2004). Interactions between phasic alerting and spatial orienting: effects of normal aging and Alzheimer’s disease. Neuropsychology, 18(2), 258–268. Forstmann, B. U., Tittgemeyer, M., Wagenmakers, E.-J., Derrfuss, J., Imperati, D., & Brown, S. (2011). The speed-accuracy tradeoff in the elderly brain: a structural model-based approach. Journal of Neuroscience, 31(47), 17242–17249. Fuxe, K. (1965). Evidence for the existence of monoamine neurons in the central nervous system. Zeitschrift Für Zellforschung Und Mikroskopische Anatomie, 65(4), 573– 596. Gabay, S., Pertzov, Y., & Henik, A. (2011). Orienting of attention, pupil size, and the norepinephrine system. Attention, Perception & Psychophysics, 73(1), 123–9. Geva, R., Zivan, M., Warsha, A., & Olchik, D. (2013). Alerting, orienting or executive attention networks: differential patters of pupil dilations. Frontiers in Behavioral Neuroscience, 7(October), 145. Gilzenrat, M. S., Nieuwenhuis, S., & Cohen, J. D. (2010). Pupil diameter tracks changes in control state predicted by the adaptive gain theory of locus coeruleus function. Cognitive, Affective, & Behavioral Neuroscience, 10(2), 252–269. Goldinger, S. D., & Papesh, M. H. (2012). Pupil Dilation Reflects the Creation and Retrieval of Memories. Current Directions in Psychological Science, 21(2), 90–95. Granholm, E., & Steinhauer, S. R. (2004). Pupillometric measures of cognitive and emotional processes. International Journal of Psychophysiology: Official Journal of the International Organization of Psychophysiology, 52(1), 1. Grudzien, A., Shaw, P., Weintraub, S., Bigio, E., Mash, D. C., & Mesulam, M. M. (2007). Locus coeruleus neurofibrillary degeneration in aging, mild cognitive impairment and early Alzheimer’s disease. Neurobiology of Aging, 28(3), 327–335. Hackley, S. A., & Valle-Inclan, F. (1998). Automatic alerting does not speed late motoric processes in a reaction-time task. Nature, 391(February), 786–788. Hämmerer, D., Callaghan, M. F., Hopkins, A., Kosciessa, J., Betts, M., Cardenas-Blanco, A., … Dolan, R. J. (2018). Locus coeruleus integrity in old age is selectively related to memories linked with salient negative events. Proceedings of the National Academy of Sciences, 115(9), 2228–2233. Hämmerer, D., Hopkins, A., Betts, M. J., Maaß, A., Dolan, R. J., & Düzel, E. (2017). Emotional arousal and recognition memory are differentially reflected in pupil diameter responses during emotional memory for negative events in younger and older adults. Neurobiology of Aging, 58, 129–139. Harada, C. N., Love, M. C. N., & Triebel, K. L. (2013). Normal cognitive aging. Clinics in Geriatric Medicine, 29(4), 737–752. Heaver, B., & Hutton, S. B. (2011). Keeping an eye on the truth? Pupil size changes associated with recognition memory. Memory, 19(4), 398–405. HESS, E. H., & POLT, J. M. (1963). Pupil Size in Relation to Mental Activity during Simple Problem-Solving. Science, 140. Hoffing, R. C., & Seitz, A. R. (2015). Pupillometry as a glimpse into the neurochemical basis of human memory encoding. Journal of Cognitive Neuroscience. Hoogendijk, W. J. G., Feenstra, M. G. P., Botterblom, M. H. A., Gilhuis, J., Sommer, I. 55 E. C., Kamphorst, W., … Swaab, D. F. (1999). Increased activity of surviving locus ceruleus neurons in Alzheimer’s disease. Annals of Neurology: Official Journal of the American Neurological Association and the Child Neurology Society, 45(1), 82– 91. Howells, F. M., Stein, D. J., & Russell, V. A. (2012). Synergistic tonic and phasic activity of the locus coeruleus norepinephrine (LC-NE) arousal system is required for optimal attentional performance. Metabolic Brain Disease, 27(3), 267–274. Hsieh, S., Wu, M., & Tang, C.-H. (2016). Adaptive strategies for the elderly in inhibiting irrelevant and conflict no-go trials while performing the go/no-go task. Frontiers in Aging Neuroscience, 7, 243. Jennings, J. M., Dagenbach, D., Engle, C. M., & Funke, L. J. (2007). Age-related changes and the attention network task: an examination of alerting, orienting, and executive function. Neuropsychology, Development, and Cognition. Section B, Aging, Neuropsychology and Cognition, 14(July 2013), 353–369. Joshi, S., Li, Y., Kalwani, R. M., & Gold, J. I. (2016). Relationships between Pupil Diameter and Neuronal Activity in the Locus Coeruleus, Colliculi, and Cingulate Cortex. Neuron, 89(1), 221–234. Just, M. A., Carpenter, P. A., & Miyake, A. (2003). Neuroindices of cognitive workload: Neuroimaging, pupillometric and event-related potential studies of brain work. Theoretical Issues in Ergonomics Science, 4(1–2), 56–88. Kafkas, A., & Montaldi, D. (2015). The pupillary response discriminates between subjective and objective familiarity and novelty. Psychophysiology, 52(10), 1305– 1316. Kety, S. S. (1972). The possible role of the adrenergic systems of the cortex in learning. Research Publications-Association for Research in Nervous and Mental Disease, 50, 376–389. Kim, M., Beversdorf, D. Q., & Heilman, K. M. (2000). Arousal response with aging: Pupillographic study. Journal of the International Neuropsychological Society, 6(3), 348–350. Krishnamurthy, K., Nassar, M. R., Sarode, S., & Gold, J. I. (2016). Adaptive, arousal- related adjustments of perceptual biases optimize perception in a dynamic environment. BioRxiv, 83766. Kristjansson, S. D., Stern, J. A., Brown, T. B., & Rohrbaugh, J. W. (2009). Detecting phasic lapses in alertness using pupillometric measures. Applied Ergonomics, 40(6), 978–986. Kubis, N., Faucheux, B. A., Ransmayr, G., Damier, P., Duyckaerts, C., Henin, D., … Agid, Y. (2000). Preservation of midbrain catecholaminergic neurons in very old human subjects. Brain, 123(2), 366–373. Langner, R., Cieslik, E. C., Behrwind, S. D., Roski, C., Caspers, S., Amunts, K., & Eickhoff, S. B. (2015). Aging and response conflict solution: behavioural and functional connectivity changes. Brain Structure and Function, 220(3), 1739–1757. Lecas, J. C. (2004). Locus coeruleus activation shortens synaptic drive while decreasing spike latency and jitter in sensorimotor cortex. Implications for neuronal integration. European Journal of Neuroscience, 19(9), 2519–2530. Li, S.-C., Lindenberger, U., & Sikström, S. (2001). Aging cognition: from neuromodulation to representation. Trends in Cognitive Sciences, 5(11), 479–486. 56 Loewenfeld, I. E., & Lowenstein, O. (1993). The pupil: Anatomy, physiology, and clinical applications (Vol. 2). Wiley-Blackwell. Luo, Y., Zhou, J., Li, M.-X., Wu, P.-F., Hu, Z.-L., Ni, L., … Wang, F. (2015). Reversal of aging-related emotional memory deficits by norepinephrine via regulating the stability of surface AMPA receptors. Aging Cell, 14(2), 170–179. MacDonald, a W., Cohen, J. D., Stenger, V. a, & Carter, C. S. (2000). Dissociating the role of the dorsolateral prefrontal and anterior cingulate cortex in cognitive control. Science, 288(5472), 1835–1838. Manaye, K. F., McIntire, D. D., Mann, D., & German, D. C. (1995). Locus coeruleus cell loss in the aging human brain: A non-random process. Journal of Comparative Neurology, 358(1), 79–87. Mann, D. M. A. (1983). The locus coeruleus and its possible role in ageing and degenerative disease of the human central nervous system. Mechanisms of Ageing and Development, 23(1), 73–94. Marcyniuk, B., Mann, D. M. A., & Yates, P. O. (1989). The topography of nerve cell loss from the locus caeruleus in elderly persons. Neurobiology of Aging, 10(1), 5–9. Mather, M., & Harley, C. W. (2016). The Locus Coeruleus: Essential for Maintaining Cognitive Function and the Aging Brain. Trends in Cognitive Sciences, 20(3), 214– 226. Mayr, U., Awh, E., & Laurey, P. (2003). Conflict adaptation effects in the absence of executive control. Nature Neuroscience, 6(5), 450–452. McDougal, D. H., & Gamlin, P. D. (2015). Autonomic control of the eye. Comprehensive Physiology, 5(1), 439. Miller, M. A., Kolb, P. E., Leverenz, J. B., Peskind, E. R., & Raskind, M. A. (1999). Preservation of noradrenergic neurons in the locus ceruleus that coexpress galanin mRNA in Alzheimer’s disease. Journal of Neurochemistry, 73, 2028–2036. Moloney, K. P., Jacko, J. A., Vidakovic, B., Sainfort, F., Leonard, V. K., & Shi, B. (2006). Leveraging data complexity: Pupillary behavior of older adults with visual impairment during HCI. ACM Transactions on Computer-Human Interaction (TOCHI), 13(3), 376–402. Mouton, P. R., Pakkenberg, B., Gundersen, H. J. G., & Price, D. L. (1994). Absolute number and size of pigmented locus coeruleus neurons in young and aged individuals. Journal of Chemical Neuroanatomy, 7(3), 185–190. Murphy, P. R., O’Connell, R. G., O’Sullivan, M., Robertson, I. H., & Balsters, J. H. (2014). Pupil diameter covaries with BOLD activity in human locus coeruleus. Human Brain Mapping, 35(8), 4140–4154. Murphy, P. R., Robertson, I. H., Balsters, J. H., & O’connell, R. G. (2011). Pupillometry and P3 index the locus coeruleus-noradrenergic arousal function in humans. Psychophysiology, 48(11), 1532–1543. Naber, M., Frässle, S., Rutishauser, U., & Einhäuser, W. (2013). Pupil size signals novelty and predicts later retrieval success for declarative memories of natural scenes. Journal of Vision, 13(2), 11. Nassar, M. R., Rumsey, K. M., Wilson, R. C., Parikh, K., Heasly, B., & Gold, J. I. (2012). Rational regulation of learning dynamics by pupil-linked arousal systems. Nature Neuroscience, 15(7), 1040–1046. Nichols, T. E., & Holmes, A. P. (2001). Nonparametric Permutation Tests for Functional 57 Neuroimaging Experiments: A Primer with examples. Human Brain Mapping, 15(1), 1–25. Nieuwenhuis, S., De Geus, E. J., & Aston-Jones, G. (2011). The anatomical and functional relationship between the P3 and autonomic components of the orienting response. Psychophysiology, 48(2), 162–175. Otero, S. C., Weekes, B. S., & Hutton, S. B. (2011). Pupil size changes during recognition memory. Psychophysiology, 48(10), 1346–1353. Papesh, M. H., Goldinger, S. D., & Hout, M. C. (2012). Memory strength and specificity revealed by pupillometry. International Journal of Psychophysiology, 83(1), 56–64. Petersen, S. E., & Posner, M. I. (2012). The Attention System of the Human Brain: 20 Years After. Annual Review of Neuroscience, 35(1), 73–89. Peysakhovich, V., Vachon, F., & Dehais, F. (2017). The impact of luminance on tonic and phasic pupillary responses to sustained cognitive load. International Journal of Psychophysiology, 112, 40–45. Piquado, T., Isaacowitz, D., & Wingfield, A. (2010). Pupillometry as a measure of cognitive effort in younger and older adults. Psychophysiology, 47(3), 560–569. Posner, M. I., Klein, R., Summers, J., & Buggie, S. (1973). On the selection of signals. Memory & Cognition, 1(1), 2–12. Posner, M. I., & Petersen, S. E. (1990). The attention system of the human brain. Annual Review of Neuroscience, 13(1), 25–42. Rajkowski, J., Kubiak, P., & Aston-Jones, G. (1994). Locus Coeruleus Activity in Monkey : Phasic and Tonic Changes Are Associated With Altered Vigilance. Brain Research Bulletin, 35, 607–616. Rajkowski, J., Majczynski, H., Clayton, E., & Aston-Jones, G. (2004). Activation of Monkey Locus Coeruleus Neurons Varies With Difficulty and Performance in a Target Detection Task. Journal of Neurophysiology, 92(1), 361–371. Reimer, J., Froudarakis, E., Cadwell, C. R., Yatsenko, D., Denfield, G. H., & Tolias, A. S. (2014). Pupil fluctuations track fast switching of cortical states during quiet wakefulness. Neuron, 84(2), 355–362. Reimer, J., Mcginley, M. J., Liu, Y., Rodenkirch, C., Wang, Q., Mccormick, D. A., & Tolias, A. S. (2016). Pupil fluctuations track rapid changes in adrenergic and cholinergic activity in cortex. Nature Publishing Group, 7(May), 1–7. Rey-Mermet, A., & Gade, M. (2017). Inhibition in aging: What is preserved? What declines? A meta-analysis. Psychonomic Bulletin & Review, 1–22. Robertson, I. H. (2013). A noradrenergic theory of cognitive reserve: Implications for Alzheimer’s disease. Neurobiology of Aging, 34(1), 298–308. Samuels, E. R., & Szabadi, E. (2008). Functional neuroanatomy of the noradrenergic locus coeruleus: its roles in the regulation of arousal and autonomic function part I: principles of functional organisation. Current Neuropharmacology, 6(3), 235–53. Sanders, A. F. (1980). 20 Stage analysis of reaction processes. Advances in Psychology, 1, 331–354. Sara, S. J. (2000). Strengthening the shaky trace through retrieval. Nature Reviews. Neuroscience, 1(3), 212–213. Sara, S. J., & Bouret, S. (2012). Orienting and Reorienting: The Locus Coeruleus Mediates Cognition through Arousal. Neuron, 76(1), 130–141. Sara, S. J., & Devauges, V. (1988). Priming stimulation of locus coeruleus facilitates 58 memory retrieval in the rat. Brain Research, 438(1–2), 299–303. Sara, S. J., & Hars, B. (2006). In memory of consolidation. Learning & Memory, 13, 515–521. Schacht, A., Dimigen, O., & Sommer, W. (2010). Emotions in cognitive conflicts are not aversive but are task specific. Cognitive, Affective & Behavioral Neuroscience, 10(3), 349–356. Scharinger, C., Soutschek, A., Schubert, T., & Gerjets, P. (2015). When flanker meets the n-back: What EEG and pupil dilation data reveal about the interplay between the two central-executive working memory functions inhibition and updating. Psychophysiology, 52(10), 1293–1304. Schwarz, L. A., & Luo, L. (2015). Organization of the locus coeruleus-norepinephrine system. Current Biology, 25(21), R1051–R1056. Shenhav, A., Botvinick, M. M., & Cohen, J. D. (2013). The expected value of control: an integrative theory of anterior cingulate cortex function. Neuron, 79(2), 217–240. Shibata, E., Sasaki, M., Tohyama, K., Kanbara, Y., Otsuka, K., Ehara, S., & Sakai, A. (2006). Age-related changes in locus ceruleus on neuromelanin magnetic resonance imaging at 3 Tesla. Magn Reson Med Sci, 5(4), 197–200. Simon, J. R., & Wolf, J. D. (1963). Choice reaction time as a function of angular stimulus-response correspondence and age. Ergonomics, 6(1), 99–105. Sirois, S., & Brisson, J. (2014). Pupillometry. Wiley Interdisciplinary Reviews: Cognitive Science, 5(6), 679–692. Snodgrass, J. G., & Vanderwart, M. (1980). A standardized set of 260 pictures: Norms for name agreement, image agreement, familiarity, and visual complexity. Journal of Experimental Psychology: Human Learning & Memory, 6(2), 174–215. Spokes, E. G. S. (1979). An analysis of factors influencing measurements of dopamine, noradrenaline, glutamate decarboxylase and choline acetylase in human post- mortem brain tissue. Brain: A Journal of Neurology, 102(2), 333–346. Stanislaw, H., & Todorov, N. (1999). Calculation of signal detection theory measures. Behavior Research Methods, Instruments, & Computers, 31(1), 137–149. Sterpenich, V., D’Argembeau, A., Desseilles, M., Balteau, E., Albouy, G., Vandewalle, G., … Maquet, P. (2006). The Locus Ceruleus Is Involved in the Successful Retrieval of Emotional Memories in Humans. Journal of Neuroscience, 26(28), 7416–7423. Sun, F. W., Stepanovic, M. R., Andreano, J., Barrett, L. F., Touroutoglou, A., & Dickerson, B. C. (2016). Youthful brains in older adults: preserved neuroanatomy in the default mode and salience networks contributes to youthful memory in superaging. Journal of Neuroscience, 36(37), 9659–9668. Takezawa, T., & Miyatani, M. (2005). Quantitative relation between conflict and response inhibition in the Flanker task. Psychological Reports, 97(2), 515–526. Tales, A., Muir, J. L., Bayer, A., & Snowden, R. J. (2002). Spatial shifts in visual attention in normal ageing and dementia of the Alzheimer type. Neuropsychologia, 40(12), 2000–2012. Teufel, H. J., & Wehrhahn, C. (2000). Evidence for the contribution of S cones to the detection of flicker brightness and red-green. J Opt Soc Am A Opt Image Sci Vis, 17(6), 994–1006. Tiplady, B., Bowness, E., Stien, L., & Drummond, G. (2005). Selective effects of 59 clonidine and temazepam on attention and memory. Journal of Psychopharmacology (Oxford, England), 19(3), 259–265. Tona, K. D., Murphy, P. R., Brown, S. B. R. E., & Nieuwenhuis, S. (2016). The accessory stimulus effect is mediated by phasic arousal: A pupillometry study. Psychophysiology, 53(7), 1108–1113. Tulving, E. (1982). Synergistic ecphory in recall and recognition. Canadian Journal of Psychology/Revue Canadienne de Psychologie, 36(2), 130. Unsworth, N., & Robison, M. K. (2016). Pupillary correlates of lapses of sustained attention. Cognitive, Affective, & Behavioral Neuroscience, (April), 601–615. van Bochove, M. E., Van der Haegen, L., Notebaert, W., & Verguts, T. (2013). Blinking predicts enhanced cognitive control. Cognitive, Affective & Behavioral Neuroscience, 13(2), 346–54. Van Den Brink, R. L., Murphy, P. R., & Nieuwenhuis, S. (2016). Pupil diameter tracks lapses of attention. PLoS ONE, 11(10), 1–16. van der Wel, P., & van Steenbergen, H. (2018). Pupil dilation as an index of effort in cognitive control tasks: A review. Psychonomic Bulletin & Review, 1–11. Van Gerven, P. W. M., Paas, F., Van Merriënboer, J. J. G., & Schmidt, H. G. (2004). Memory load and the cognitive pupillary response in aging. Psychophysiology, 41(2), 167–174. van Rijn, H., Dalenberg, J. R., Borst, J. P., & Sprenger, S. A. (2012). Pupil Dilation Co- Varies with Memory Strength of Individual Traces in a Delayed Response Paired- Associate Task. PLoS ONE, 7(12). van Steenbergen, H., & Band, G. P. H. (2013). Pupil dilation in the Simon task as a marker of conflict processing. Frontiers in Human Neuroscience, 7(May), 215. Varazzani, C., San-Galli, A., Gilardeau, S., & Bouret, S. (2015). Noradrenaline and dopamine neurons in the reward/effort trade-off: a direct electrophysiological comparison in behaving monkeys. Journal of Neuroscience, 35(20), 7866–7877. Vijayashankar, N., & Brody, H. (1979). A quantitative study of the pigmented neurons in the nuclei locus coeruleus and subcoeruleus in man as related to aging. Journal of Neuropathology & Experimental Neurology, 38(5), 490–497. Võ, M. L. H., Jacobs, A. M., Kuchinke, L., Hofmann, M., Conrad, M., Schacht, A., & Hutzler, F. (2008). The coupling of emotion and cognition in the eye: Introducing the pupil old/new effect. Psychophysiology, 45(1), 130–140. Wheeler, M. E., Petersen, S. E., & Buckner, R. L. (2000). Memory’s echo: vivid remembering reactivates sensory-specific cortex. Proceedings of the National Academy of Sciences, 97(20), 11125–11129. Wilson, R. S., Nag, S., Boyle, P. A., Hizel, L. P., Yu, L., Buchman, A. S., … Bennett, D. A. (2013). Neural reserve, neuronal density in the locus ceruleus, and cognitive decline. Neurology, 80(13), 1202–1208. Ziaei, M., von Hippel, W., Henry, J. D., & Becker, S. I. (2015). Are age effects in positivity influenced by the valence of distractors? PloS One, 10(9), e0137604. 60 Tables Table 1. Demographics and Neuropsychological Testing Summary YC (n=25) ED (n=25) Mean STD Mean STD Age** 20.0 2.2 74.1 6.4 Sex (Male Female) 6 19 8 17 Years of Education** 14.0 2.1 15.7 2.4 WIDS - forward 12.2 2.2 12.2 2.3 WIDS - back 10.5 2.1 9.8 2.1 WIDS - sequence** 10.4 1.6 9.1 1.3 DKVF - letter 47.0 9.0 46.0 10.3 DKVF - category 45.0 4.1 45.3 9.5 DKVF - category switch** 16.1 2.8 13.4 3.1 DKVF - switch accuracy** 15.4 2.9 12.7 3.0 Trails A** 23.2 7.4 37.4 9.0 Trails B** 45.1 11.9 80.2 30.2 VSAT Total 36.8 28.4 34.6 21.6 PSQI 3.8 2.0 3.4 2.2 ANART -- -- 121.1 4.3 MMSE -- -- 28.9 1.0 RBANS – total scaled index -- -- 111.5 11.5 CRIq -- -- 133.9 16.1 ** = p < .01 indicate significant aging effects on the measure. Abbreviations are: WIDS = Wechsler Adult Intelligence Scale IV Digit Span Task; DKVF = Delis-Kaplan Executive Function System Verbal Fluency Task; VSAT = Visual Search and Attention Test; PSQI = Pittsburgh Sleep Quality Index; ANART = American National Adult Reading Test; MMSE = Mini-Mental State Examination; RBANS = Repeatable Battery for the Assessment of Neuropsychological Status; CRIq = Cognitive Reserve Index questionnaire. 61 Table 2. Mean and standard deviation of accuracies and RTs in all tasks YC (n=25) ED (n=25) Mean STD Mean STD Phasic Alerting Task Mean Accuracy 0.9948 0.0074 0.9968 0.0051 Response Time • Sound Condition** 292.03 18.85 404.56 79.71 • No Sound Condition** 318.37 27.08 457.43 91.70 Paired sample STD 18.72 27.33 Recognition Memory Task Mean Accuracy** 0.8558 0.0693 0.7900 0.0718 Hit Rate** 0.8112 0.1063 0.6880 0.1711 False Alarm Rate 0.1992 0.1135 0.2160 0.1536 A’* 0.8745 0.0661 0.8282 0.0662 B’’ -0.0033 0.3561 0.1524 0.4538 Flanker Task Mean Accuracy 0.9830 0.0117 0.9830 0.0141 Response Time • Congruent Condition** 366.77 36.34 510.16 80.68 • Incongruent Condition** 397.29 35.57 532.80 81.93 Paired sample STD 13.11 17.07 * = p < .05, ** = p < .01 indicate significant aging effects on the variables. 62 Figures Figure 1. Experiment flow chart for Phasic Alerting Task; actual presentation colors are isoluminant Figure 2. Experiment flow chart for Recognition Memory Task; actual presentation colors are isoluminant 63 Figure 3. Experiment flow chart for Flanker Task; actual presentation colors are isoluminant 64 Figure 4. Mean RTs of Sound and No Sound conditions in Phasic Alerting Task, including only correct trials. Errors bars indicate standard errors. Figure 5. Raw pupil dilation (PD) waveforms showing the effect of the auditory tone on PD computed as subtracting the mean of baseline from trial pupil diameters. 65 Figure 6. Plot of intercept terms capturing sizes of trial pupil diameter in the Phasic Alerting Task. 66 Figure 7. Temporal plots (2s) of pupillary responses in Phasic Alerting Task: shaded areas represent standard error from mean. Time points when one-sample t-tests of B values resulted in p < 0.05 significance level are marked red. The time points with the lowest p values are marked with dotted red lines. 67 Figure 8. Temporal plots (2s) of speed effects modeled separately in sound and no sound trials in (top) YC group and (bottom) ED group. 68 Figure 9. Hit rates and correct rejection rates in Recognition Memory Task. Dotted lines indicate the mean hit and rejection rates in each age group. Figure 10. Raw PD waveforms showing the effect of old/new words on PD computed as subtracting the mean of baseline from trial pupil diameters; only correct trials are included to generate this diagram. 69 Figure 11. Temporal plots (2s) of pupillary responses in Recognition Memory Task: shaded areas represent standard error from mean. Time points when one-sample t- tests of B values resulted in p < 0.05 significance level are marked red. The time points with the lowest p values are marked with dotted red lines. 70 Figure 12. Temporal plots (2s) of pupillary old/new effects modeled separately in correct and incorrect trials in (top) YC group and (bottom) ED group. 71 Figure 13. Mean RTs of Congruent and Incongruent conditions in Flanker Task, including only correct trials. Errors bars indicate standard errors. Figure 14. Raw PD waveforms showing the effect of congruency conditions on PD computed as subtracting the mean of baseline from trial pupil diameters 72 Figure 15. Temporal plots (2s) of pupillary responses in Flanker Task: shaded areas represent standard error from mean. Time points when one-sample t-tests of B values resulted in p < 0.05 significance level are marked red. The time points with the lowest p values are marked with dotted red lines. 73