How arousal-related neurotransmitter systems compensate for age-related decline Mara Mather University of Southern California Index terms: Aging, arousal, neurotransmitters, norepinephrine, noradrenaline, dopamine, acetylcholine, orexin, histamine, homeostasis, locus coeruleus. Citation: Mather, M., (in press). How arousal-related neurotransmitter systems compensate for age- related decline. In Gutchess, A. & Thomas, A., ed. The Cambridge Handbook of Cognitive Aging: A Life Course Perspective. Cambridge University Press Abstract Without brain systems that modulate arousal, we would not be able to have daily sleep-wake cycles, focus attention when needed, experience emotional responses, or even maintain consciousness. Thus, it is not surprising that there are multiple overlapping neurotransmitter systems that control arousal. In aging, most of these systems show decline in basic features such as number of receptors, transporters, and sometimes even in neuron count. These declines have the potential to lead to significant problems in maintaining basic arousal functions. However, some of these systems also show compensatory increases that allow for maintained levels of circulating neurotransmitters in the system—but at the cost of reduced dynamic range in arousal responses. 2 How arousal-related neurotransmitter systems compensate for age-related decline In the 1960’s, opposing theories were proposed about arousal and aging. On the one hand, “overarousal” theory (Eisdorfer, 1968) proposed that older adults’ learning and performance deficits were due to their overarousal during laboratory tasks. In contrast, “underarousal” theory (Birren, 1960; Birren, Cunningham, & Yamamoto, 1983; Falk & Kline, 1978) argued that with age, performance deficits were due to a decreased baseline activation level and lessened reactivity of the central nervous system in older adults. One might think that, fifty years later, this debate would either be resolved or irrelevant. However, decades of accumulated findings regarding arousal systems in the brain continue to provide data to support both perspectives. In this chapter, I outline the current cases for both underarousal and overarousal in aging. I then make the case that both processes occur simultaneously in aging. Most neurotransmitter systems involved in arousal show significant decline in aging, which could lead to chronic underarousal. However, increased tonic levels of some arousal-related neurotransmitters compensates for these declines. This compensation comes at the cost of a reduced dynamic range of responses. 1. What Is Arousal? Arousal is a general term that can be broadly categorized as covering three domains: wakefulness, autonomic arousal, and affective arousal (Satpute, Kragel, Barrett, Wager, & Bianciardi, 2018). Brainstem nuclei and the hypothalamus, thalamus, posterior cingulate cortex, precuneus and medial prefrontal cortex have been implicated in all three types of arousal. Autonomic and affective arousal involve additional regions including the amygdala, the insula and anterior cingulate cortex (Satpute et al., 2018). Within the brainstem, five neurotransmitter systems—norepinephrine, dopamine, serotonin, acetylcholine, and histamine—have overlapping roles in increasing arousal (Pfaff, 2006), as does a neuropeptide, orexin. These fate of these systems in aging and the implications for arousal are the focus of this chapter. I first review each of these briefly below. 1.1 Norepinephrine The locus coeruleus, a nucleus in the pons, is the source of most of the brain’s norepinephrine. It receives direct inputs from all of the other arousal systems outlined below and serves as a hub region to integrate all categories of arousal (Mather, in press). Its tonic levels of activity are associated with wakefulness levels and its phasic levels closely track momentary fluctuations in arousal induced by all sorts of conditions, including emotional arousal, cognitive or physical effort, and detection of salient stimuli. Locus coeruleus neurons have long axons that release norepinephrine throughout much of the brain. Typically, the release occurs at bulges along the axons known as varicosities, rather than at synapses. Thus, rather than having a specific post-synaptic target, norepinephrine is released into extra- synaptic space where it can have a broader impact. 1.2 Dopamine Dopamine modulates behavioral motivation, playing central roles in both reward and movement (Cools, Nakamura, & Daw, 2011; Volkow, Wise, & Baler, 2017). Many motivated behaviors, such as eating, copulating, or taking addictive drugs, require activity and attention to be implemented. Thus, it is not surprising that the dopamine system interacts with the other arousal systems to modulate arousal levels, in order to increase response vigor when higher rewards are at stake (Niv, Daw, Joel, & Dayan, 2007). Dopaminergic signaling in the ventral 3 tegmental area is correlated with vigilance levels and sleep/wake timing (Wisor, 2018). Dopamine neurons in the dorsal raphe are activated by salient stimuli and modulate wakefulness (Cho et al., 2017). 1.3 Serotonin In some ways, serotonin plays an opposing role to dopamine, by modulating aversive processing and behavioral inhibition (Boureau & Dayan, 2011; Cools et al., 2011). Thus, like dopamine, serotonin is linked with arousal via its role in motivated behavior. In addition, the predominant characteristic of most serotonin neurons in the raphe nucleus is that their activity is related to the sleep-wake cycle, although whether activity is associated with sleep or wake states depends on the current behavioral state and other factors (Ursin, 2002). 1.4 Acetylcholine Acetylcholine plays a role in selective attention, orienting and detecting behaviorally significant stimuli (Klinkenberg, Sambeth, & Blokland, 2011). Cholinergic neurons in the basal forebrain are most active during wakefulness, and the levels of their activity correlate with cortical activation (for review see Tyree & de Lecea, 2017). Acetylcholine neurons use GABA as a co-transmitter, which may increase the signal-to-noise ratio in sensory signals and increase control over cortical plasticity (Ma, Hangya, Leonard, Wisden, & Gundlach, 2018). 1.5 Histamine Histamine, released broadly throughout the brain from the tuberomamillary nucleus of the hypothalamus, is responsible for modulating wakefulness and consciousness levels, and is often a target of drugs used for anesthesia (Haas & Panula, 2003; Wada, Inagaki, Yamatodani, & Watanabe, 1991). Like norepinephrine, histamine is primarily released via varicosities into extracellular space rather than at specific synapses (Takagi et al., 1986). In addition to releasing histamine, histamine neurons release GABA into the extrasynaptic space, which seems to play an important role in calibrating the effects of histamine release and avoiding overarousal (Yu et al., 2015). 1.6 Orexin In addition, orexin (or hypocretin), a neuropeptide synthesized in the lateral hypothalamus, also plays an important role in modulating arousal, promoting stable periods of wakefulness and sustained alertness needed during motivated behavior (Alexandre, Andermann, & Scammell, 2013), as well as promoting award-based feeding (Cason et al., 2010). Orexin neurons have reciprocal connections with all of the nuclei discussed above: the locus coeruleus (NE), tuberomammillary nucleus (histamine), dorsal raphe (serotonin), ventral tegmental area and nucleus accumbens (dopamine), basal forebrain and laterodorsal and pedunculopontine tegmental nuclei (acetylcholine) (Alexandre et al., 2013). Genetic mutations disrupting orexin lead to narcolepsy, or the inability to sustain long periods of wakefulness. However, microdialysis probes in the amygdala of human epilepsy patients revealed that orexin levels are not a simple function of arousal, as they are highest during positive emotion and social interactions, somewhat elevated during anger, and lowest during episodes of waking pain and during sleep (Blouin et al., 2013). These findings suggest that orexin promotes approach-related arousal (associated with positive emotion and anger) rather than arousal in general. 1.7 Summary 4 This brief overview underscores the multifaceted nature of arousal. Throughout each day and night, multiple neurotransmitters interact to induce and maintain sleep and alertness. These neurotransmitters have overlapping influences but also each have their own signature pattern of effects. As reviewed in the next section, there is evidence for age-related decline in all of these systems (e.g., Rieckmann & Nyberg, in press) that could support a characterization of aging as being associated with underarousal. However, as addressed in the section after that, the story is not that simple. 2. Evidence for Underarousal in the Aging Brain 2.1 Multiple Indicators of Decline in the Locus Coeruleus-Norepinephrine (LC-NE) System Although modern unbiased stereology techniques (i.e., use of random sampling to count neurons within two-dimensional tissue samples to extract estimates of the count within a three-dimensional area) and do not tend to find associations between age and post-mortem count of LC neurons (Mouton, Pakkenberg, Gundersen, & Price, 1994; Ohm, Busch, & Bohl, 1997; Theofilas et al., 2017), there are other indicators of decline within the LC-NE system (Mather & Harley, 2016). In particular, tau pathology increases with age in the LC (Braak, Thal, Ghebremedhin, & Del Tredici, 2011). In healthy neurons, tau protein is mostly found in axons, where it binds to tubulin, helping to stabilize and stiffen microtubules which help support the extended axon structure and transport within the axon (Arendt, Stieler, & Holzer, 2016). Hyperphosphorylation of tau reduces its binding to tubulin and is an initial phase before eventually becoming aggregated as seen in Alzheimer’s disease and other pathological conditions (Iqbal, Liu, & Gong, 2016). Initial signs of hyperphosphorylated tau are seen quite early in life in the locus coeruleus, and appear to slowly spread from the brainstem to entorhinal cortex and then other cortical regions in healthy aging, with this process accelerated in those exhibiting signs of Alzheimer’s disease (Braak et al., 2011). It is not clear yet how hyperphosphorylated tau affects neuronal function in the absence of other pathology. Given its key role in microtubule stabilization, it may reduce axonal transport effectiveness. In addition, recent findings indicate that small amounts of tau in the dendritic compartment of neurons support both NMDA and AMPA receptor function and so increasing tau phosphorylation reduces signal transduction and prevents excitotoxicity which otherwise could occur from overexcitation of NMDA receptors (Arendt et al., 2016). Another indication that the LC-NE system declines in aging is that NE levels are lower in older than younger adults found in sections of brain tissue from sites such as the cingulate gyrus (Arranz et al., 1996; Winblad, Hardy, Bäckman, & Nilsson, 1985), hippocampus (Winblad et al., 1985), hypothalamus (Winblad et al., 1985), and hindbrain (Robinson, 1975). One qualifying point that I will return to in a later section, however, is that in these studies, the tissue is blended into a solution that does not allow investigators to discriminate NE levels found in vesicles (storage sites) within a neuron from those in current circulation available to influence activity. 2.2. Dopamine System Decreases With Age Positron emission tomography studies indicate decline in D , D , and D dopamine 1 2 3 receptors and in dopamine transporters (for review see Bäckman, Lindenberger, Li, & Nyberg, 2010). Post mortem studies also indicate loss of dopaminergic neurons in the substantia nigra with age (Fearnley & Lees, 1991; McGeer, McGeer, & Suzuki, 1977) and significant decline in dopamine levels in striatum (Carlsson & Winblad, 1976; Kish, Shannak, Rajput, Deck, & Hornykiewicz, 1992), hippocampus (Adolfsson, Gottfries, Roos, & Winblad, 1979) and hypothalamus (Arranz et al., 1996). 5 2.3 Serotonin Receptors Show Decline Post-mortem studies reveal reduced density of serotonin receptors in the brain while PET studies show reduced receptor binding (for reviews see Meltzer et al., 1998; Rodríguez, Noristani, & Verkhratsky, 2012). 2.4 Brain Orexin Levels Decrease Postmortem analysis of the hypothalamus reveals a 10% decline in orexin neurons from young adulthood (between 22 and 32) and later adulthood (between age 48 and 60) (Hunt, Rodriguez, Waters, & Machaalani, 2015). In a study with male rhesus monkeys, orexin neuron numbers did not differ with age, but there was less excitatory orexin innervation to the locus coeruleus (Downs et al., 2007). 2.5 Histamine Binding Decreases PET studies show decreases in histamine receptor binding (specifically, H R-binding) 1 with age throughout much of the cortex (Higuchi et al., 2000; Yanai et al., 1992). 2.6 Acetylcholine System Shows Less Age-Related Change In the 1970’s and 80’s, the acetylcholine hypothesis of Alzheimer’s disease emerged based on finding significant loss of basal forebrain cholinergic neurons in brains affected by Alzheimer’s disease (Davies & Maloney, 1976; Perry, Gibson, Blessed, Perry, & Tomlinson, 1977; Whitehouse et al., 1982). However, more recent evidence indicates that notable cholinergic system deficits do not emerge until late stages of Alzheimer’s disease (Davis, Mohs, Marin, & et al., 1999; Gilmor et al., 1999; Schliebs & Arendt, 2011) and aspects of the system such as number of cholinergic neurons in the basal forebrain (Gilmor et al., 1999) and nicotinic acetylcholine receptors (Jogeshwar et al., 2018) are not correlated with age (although the volume of the basal forebrain cholinergic system declines throughout adulthood; Grothe, Heinsen, & Teipel, 2012). However, a significant risk in late life for this system comes from anticholinergic medications prescribed for a variety of common ailments (e.g., depression, allergies, cold and flu symptoms, insomnia, and urinary incontinence). Estimates of how many older adults take anticholinergic medication is as high as 37% (Britt & Day, 2016), and higher cumulative use of anticholinergics is associated with greater risk of developing Alzheimer’s disease (Gray et al., 2015). Consistent with acetylcholine’s role in arousal, cholinesterase inhibitors that increase acetylcholine levels improve attention and concentration and reduce anxiety and restlessness in patients experiencing those symptoms (Lemstra, Eikelenboom, & van Gool, 2003). 3 Evidence for Overarousal in Aging The previous section reviewed an array of declines in arousal-related transmitter systems. This section presents evidence that such declines are associated with increases in transmitter activity that may help maintain function, especially within the hub arousal system, the LC-NE system. 3.1 The LC-NE System Shows Compensatory Increases in Activity As outlined in the previous section, there is evidence of decline in the LC-NE system in aging and studies that examine NE levels within homogenized (i.e., blended into a consistent solution) brain tissue find decreases in older adults. However, this quantification of overall brain levels of NE may only tell part of the story, as there is also evidence of LC-NE hyperactivity in 6 aging and in early stages of the AD process (Gannon & Wang, 2018; Weinshenker, 2018). Norepinephrine levels in cerebrospinal fluid and in blood are greater in older adults than in younger adults (Elrod et al., 1997; Raskind et al., 1988; White et al., 1997). For instance, one estimate is that plasma norepinephrine increases 10-15% in concentration per decade during adulthood (Seals & Esler, 2000). The increases in norepinephrine in cerebrospinal fluid are even greater in patients with Alzheimer’s disease than in healthy older adults (Elrod et al., 1997). While these estimates can be confounded by age differences in the rate of flushing out waste from cerebrospinal fluid, techniques that assess the rate of norepinephrine spillover into plasma find elevated rates in older compared with younger adults (Seals & Esler, 2000). In addition, muscle sympathetic nerve activity increases as people age (Fagius & Wallin, 1993), and in general, increasing sympathetic nerve activity seems to contribute to increasing blood pressure with age (Hart & Charkoudian, 2014). Why would LC-NE system activity increase in aging? One possibility is that as LC neurons die in aging and Alzheimer’s disease, surviving LC neurons go into overdrive to compensate for this loss (Gannon & Wang, 2018; Weinshenker, 2018). An impressive demonstration of the ability of surviving LC neurons to compensate for the loss of other LC neurons comes from a study that induced partial LC lesions in rats and then measured extracellular NE levels in the hippocampus using microdialysis both at baseline and during stress (Abercrombie & Zigmond, 1989). Reductions of hippocampal tissue NE (detected from homogenized postmortem tissue) of up to 50% had no effect on NE levels detected in extracellular space; it took more than 50% decline in tissue NE to see a decline in NE detected using microdialysis. These findings suggest that remaining LC neurons increased their basal rate of NE release to compensate for declines in total LC neurons. Furthermore, destroying most of the noradrenaline terminals in the forebrain of rats led the intact LC neurons to increase firing rates for the next few weeks at about a 4-fold rate (Chiodo, Acheson, Zigmond, & Stricker, 1983) and destroying terminals in the hippocampus lead to a rapid increase in the activity of the rate-limiting enzyme for NE production, tyrosine hydroxylase (Acheson, Zigmond, & Stricker, 1980; Acheson & Zigmond, 1981). Destruction of LC neurons can also lead to increases in NE axons in forebrain regions over the next 12 months (Fritschy & Grzanna, 1992). Most importantly, these results indicate that even if NE levels quantified from postmortem homogenized tissue are diminished, this does not mean that extracellular circulating levels of NE are diminished. Consistent with these experimental findings in rats, patients with neuronal loss due to Alzheimer’s or dementia with Lewy bodies showed changes in the noradrenergic system consistent with compensation, including an increase in tyrosine hydroxylase mRNA expression in remaining neurons suggesting these neurons are compensating for the loss of other LC neurons by increasing the rate-limiting enzyme in the synthesis of NE (Szot et al., 2000, 2006). In addition, those with more loss of pigmented LC neurons showed higher ratios of the norepinephrine metabolite 3-methoxy-4-hydroxyphenylglycol (MHPG) to total norepinephrine levels in frontal medial cortex (Hoogendijk et al., 1999). In the same study, no such upregulation of serotonin nor dopamine metabolites was seen in the Alzheimer’s patients. In the previous section, I mentioned decreased resting pupil size among older adults as a potential indicator of diminished levels of LC-NE system tonic activity. However, while phasic LC activity increases pupil dilation, there is a paradoxical reverse effect for tonic activity, in which higher tonic levels of NE decrease pupil size (Koss, 1986). This evidence comes from studies observing that a2a noradrenergic agonists decrease pupil dilation (Clifford, Day, & Orwin, 1982; Phillips, Szabadi, & Bradshaw, 2000). Phasic pupil responses are associated with LC activity (Joshi, Li, Kalwani, & Gold, 2015), and interestingly, in middle-aged men, reduced phasic pupil responses during working 7 memory were associated with greater tonic low frequency BOLD variance during a resting- state fMRI scan in key nodes of the ventral attention network (Elman et al., 2017). Greater tonic activity in ventral attention network in older adults may also help explain why arousing circumstances such as expecting an electric shock are less effective at increasing ventral attention network activity and less likely to increase coordination with LC activity than seen in younger adults (Lee et al., 2018). Why might the LC-NE system show more upregulation of activity upon loss of neurons? The NE system may be ideally set up to allow for surviving neurons to compensate for the loss of other neurons or for upregulation of release of NE from surviving varicosities when other release sites are damaged. This is because most NE release does not occur at specific synapses, but instead into the general extracellular space. If a specific synapse is damaged it is difficult for upregulation of another synapse to compensate for this, but because NE can use a “hormonal” mode of modulating activity, releasing more NE from nearby varicosities should be able to provide some compensation for loss of neighboring NE release points. It has also been proposed that even before there is neuron loss in the LC-NE system, increasing tau pathology triggers changes that lead to hyperactivity, which may help compensate for the loss of noradrenergic axons, terminals, and NE in brain regions to which the LC projects (Weinshenker, 2018). 3.2 Less Capacity to Increase Sympathetic Activity In general, compared with younger adults, stress and arousal manipulations show either similar or reduced increases in markers of sympathetic activity in older adults. For instance, in response to postural stimulation, younger adults show greater increases in NE levels, heart rate, heart rate variability and salivary alpha amylase (Lavi et al., 2007; Lipsitz, Mietus, Moody, & Goldberger, 1990; White et al., 1997; Yo et al., 1994). Salivary alpha amylase is another biomarker for NE, as sympathetic activity stimulates salivary glands to release this digestive enzyme (Nater & Rohleder, 2009). When several samples are collected each day to capture a diurnal profile, older adults have an overall greater daily output of salivary alpha amylase than younger adults (Birditt, Tighe, Nevitt, & Zarit, 2017; Nater, Hoppmann, & Scott, 2013; Strahler, Berndt, Kirschbaum, & Rohleder, 2010), findings that seem consistent with the possibility of baseline NE hyperactivity reviewed in the previous section. In general, negative social interactions are associated with higher sustained levels of alpha amylase during that day, an effect that did not differ by age group (Birditt et al., 2017). Likewise, in another study with both a psychosocial stress and a control group, older adults had greater salivary alpha amylase release than younger adults but there was no age difference in how much the stress task affected participants compared to the control group (Almela et al., 2011). 3.3 LC-NE Autoregulation May Have a Higher Setpoint a2a noradrenergic receptors are the most common noradrenergic receptors in the brain and they require less NE to activate than a1 or b-adrenergic receptors. They typically are inhibitory and often serve as autoreceptors right at the sites of NE release. Thus, their activity is critical for keeping tonic levels of NE low. Across studies, a2 antagonists have more of an effect on NE activity in older adults than younger adults (Peskind, Wingerson, Murray, & et al., 1995; Raskind, Peskind, Holmes, & Goldstein, 1999), whereas a2 agonists have more of an effect on younger adults (Raskind et al., 1988). This suggests that a2 receptors are tonically more occupied in older adults, which may result in fewer remaining a2 receptors available to inhibit NE activity. Likewise, administration of NE to the LC, which usually decreases LC firing due to a2a inhibitory noradrenergic receptors, was significantly less effective at reducing firing in surviving 8 LC neurons with upregulated activity after most other LC neurons had been damaged (Chiodo et al., 1983). This indicates that some types of damage lead the LC to transition to a higher tonic mode of activity and to become less sensitive to auto-regulation feedback mechanisms to adjust its activity levels. 3.4 Dopamine metabolism may be upregulated in aging There are emerging hints that, as seen in the LC-NE system, decline within the dopamine system is not uniform and that as much of the system declines, there may be upregulation of dopamine synthesis by remaining neurons to maintain dopamine levels. CSF levels of a major dopamine metabolite, homovanillic acid, increase in aging (Brewerton, Putnam, Lewine, & Risch, 2018; Gottfries et al., 1971). It could be that homovanillic acid remains in CSF for longer in older adults than in younger adults, such that it does not actually reflect greater dopamine metabolism in the brain (e.g., Rapoport, Schapiro, & May, 2004). However, a meta-analysis of positron emission tomography studies found that, across studies, with age there was no significant decline in dopamine synthesis—despite declines in dopamine transporters and receptors (Karrer, Josef, Mata, Morris, & Samanez-Larkin, 2017). The lack of age effect in dopamine synthesis may reflect an initial upregulation that eventually declines. Current models posit that some aspects of age-related impairments in episodic memory may be caused by reduced dopamine signaling (Bäckman et al., 2010; Li & Rieckmann, 2014). However, while a study that involved administering both a dopamine agonist and a dopamine antagonist found that older adults were more sensitive to the dopaminergic perturbation than younger adults, the dopamine agonist improved memory function only in those older adults with poor baseline memory whereas the dopamine antagonist improved memory in the better performing older adults (Abdulrahman, Fletcher, Bullmore, & Morcom, 2017). This raises the possibility that in higher functioning older adults, there is some degree of overexcitation within the dopamine system. 3.5 Serotonin Metabolism May Increase CSF levels of a major serotonergic metabolite, 5-hydroxyindoleacetic acid (5-HIAA), are higher in healthy older adults than in younger adults (Brewerton et al., 2018; Gottfries et al., 1971; Yoon et al., 2017), even though they also show decreased levels in patients with Alzheimer’s disease. 3.6 Blood Levels of Orexin Increase In contrast with the age-related declines in orexin neurons, plasma levels of orexin are positively associated with age (El‐Sedeek, Korish, & Deef, 2010; Matsumura et al., 2002). In addition, patients with Alzheimer’s disease or mild cognitive impairment show higher CSF levels of orexin-A than controls (Gabelle et al., 2017; Liguori et al., 2016), and these higher levels are associated with disrupted sleep (Liguori et al., 2016). This may reflect greater production of the peptide to overcome reduced sensitivity associated with loss of orexin neurons, as in both rodents and humans, more than half of orexin neurons must be lost before significant decreases in CSF levels of orexin are detected (for review see Nixon et al., 2015). Thus, there may be an initial phase of upregulation of orexin-A associated with disrupted sleep followed by depleted levels of orexin-A in later stages of Alzheimer’s disease, leading to excessive daytime sleepiness (Fronczek et al., 2012). 3.7 CSF levels of histamine metabolites increase Mean levels of histamine metabolites are higher in CSF of older adults than in younger adults (Motawaj, Peoc'h, Callebert, & Arrang, 2010; Prell, Khandelwal, Burns, LeWitt, & Green, 9 1990). Furthermore, in Alzheimer’s disease, the tuberomammillary nucleus (the origin of histamine neurons) shows marked degeneration, yet the main histamine metabolite is only slightly (Motawaj et al., 2010) or not at all (Gabelle et al., 2017) decreased in the CSF of Alzheimer’s patients relative to controls. This suggests compensatory activation of remaining neurons. Furthermore, histamine levels in CSF were correlated with insomnia severity in patients with mild cognitive impairment and Alzheimer’s disease, suggesting that compensatory histamine activity during this disease process leads to overarousal (Gabelle et al., 2017). 3.8 Summary A number of neurotransmitter systems show signs of overarousal in aging insofar as their levels are increased in cerebrospinal fluid. In textbooks, neurotransmitter release is pictured as occurring at a synaptic gap. However, arousal-related neurotransmitters often are released into the extracellular space of the brain outside of specific synapses. CSF can provide a proxy measure of the level of these circulating neurotransmitters and their metabolites because interstitial fluid diffuses easily into the CSF (Spector, Robert Snodgrass, & Johanson, 2015). While cerebrospinal fluid is one step removed from the brain, it provides an important piece of data that typical postmortem data does not. Typically, postmortem examination of neurotransmitter levels is done by taking brain tissue and blending it up. This mixes up the neurons and the extracellular fluid (known as interstitial fluid). However, if in aging there are fewer neurons containing a particular neurotransmitter in stored vesicles but upregulation of release of that neurotransmitter, this could lead to decreased total neurotransmitter sequestered within neurons but equivalent levels or even increased neurotransmitter in the extracellular space. The CSF “overarousal” indications, however, require confirmation from other measures because age-related changes in CSF transit processes could potentially be the source of these differences. For instance, active transport processes help move HVA and 5-HIAA out of the brain and these transport processes decline in Alzheimer’s disease and also somewhat in aging (Spector et al., 2015). 4 Conclusions Modulating arousal is essential for so many everyday life functions, including consciousness. It is likely because of its essential nature that there are multiple overlapping neurotransmitter systems that help control arousal levels. In aging, there is clearly decline in these systems. Returning to the old debate introduced at the outset of this chapter as to whether aging involves overarousal or underarousal, the findings reviewed in this chapter suggest that loss of neurons, receptors and other infrastructure involved in brain arousal systems threaten to create under-arousal. However, as I argue in this chapter, these declines are compensated for by increased tonic levels of some of the involved neurotransmitters, which allows for maintained function at the cost of a diminished dynamic range. To maintain conscious awareness and cognitive functioning, in older adults, each neuron involved in these arousal-related systems appears to have to produce more neurotransmitters than it would in a younger brain. With these higher typical baseline levels, there may be less capacity to increase levels further. This picture fits with a recent model of emotional well-being in aging called ‘Strength and Vulnerability Integration’ or SAVI (Charles, 2010), which posits that physiological flexibility decreases in aging. However, SAVI also posits that under stress, older adults will show greater and more prolonged physiological stress responses, whereas the evidence reviewed here 10