In the Light of Evolution: Volume VII: The Human Mental Machinery (2014)

9

Making Lasting Memories: Remembering the Significant

JAMES L. MCGAUGH

_____________

Department of Neurobiology and Behavior, Center for the Neurobiology of Learning and Memory, University of California, Irvine, CA 92697-3800. E-mail: jlmcgaug@uci.edu.

The ability to learn and remember is essential for our survival. Remembering what has happened enables us to predict what is likely to happen and alter our behavior accordingly. As noted by Bernecker (2010, p. 1), “[r]emembering is a fundamental cognitive process, subserving virtually all other important cognitive functions . . . . Since without memory one couldn’t think, some philosophers go as far as to claim that memory is the mark of being human.” This latter claim is, of course, off of the mark, because most, if not all, animals display memory of their experiences. However, the many moments of their lives and our lives are not given equal weight in memory: we do not remember equally well all of our experiences. As James (1890, p. 643) commented, “[o]f some [experiences] no memory survives the instance of their passage. . . . Others . . . may be recalled as long as life endures. How can we explain these differences?” There are many possible explanations. Experiences that we attend to are, of course, more likely to be remembered. Some new experiences become lasting, because they fit well with and can be readily processed and integrated with existing memories (Craik and Lockhart, 1972). Additionally, beginning with the pioneering studies of Ebbinghaus (1885), we learned that memories are strengthened by repetition or retrieval (Roediger and Butler, 2011).

EMOTIONAL AROUSAL AND LASTING MEMORY

There is also extensive evidence that experiences that are emotionally arousing are well remembered (Brown and Kulik, 1977; Conway, 1995; McGaugh, 2003; Reisberg and Hertel, 2003). Experiences of unpleasant occasions, such as an automobile accident, a mugging, or learning about the death of a loved one, are remembered better than those experiences of a routine day (Stratton, 1919; Pillemer, 1984; Bohannon, 1988; Christianson, 1992; Conway et al., 1994; Conway, 1995; Neisser et al., 1996; Schmolck et al., 2000; Sharot et al., 2007). Memories of pleasant occasions, such as birthdays, holidays, and weddings, are also well retained. The strength of memories of events varies with the emotional significance of the events. Individuals who were close to San Francisco at the time of the 1989 San Francisco earthquake had better memories of the earthquake months later compared with individuals in Atlanta, Georgia (Neisser et al., 1996). Three years after the terrorist attack on September 11, 2001, individuals who were in downtown Manhattan at the time of the attack had more detailed

memories of the attack compared with individuals who were in midtown Manhattan, several miles from the attack (Sharot et al., 2007).

MODULATION OF MEMORY CONSOLIDATION

These findings clearly support Bacon’s (2000) assertion that “[m]emory is assisted by anything that makes an impression on a powerful passion, inspiring fear, for example or wonder, shame or joy.” However, such evidence provides only the beginnings of an answer to William James’ wondering about why some memories are lasting. A more comprehensive answer requires an understanding of the effects of emotional arousal that regulate the strength of memories.

Lasting memories are not created at the time of an experience. There is considerable evidence supporting the hypothesis of Mueller and Pilzecker (1900) that experiences initiate neural processes that perseverate and induce, over time, the consolidation of memory. Subsequently, Hebb (1949) proposed a dual-trace hypothesis of memory formation. According to this hypothesis, memories are initially based on the reverberation of neural circuits, and long-term memory results from synaptic changes induced by the neural reverberation. Thus, for both the consolidation hypothesis and the dual-trace hypothesis, lasting memory is formed after an experience.

The time-dependent process of memory consolidation thus provides an opportunity for conditions occurring after learning (i.e., during the consolidation of memory) to regulate the strength of memory. Studies of the effects of electroconvulsive shock (Duncan, 1949) were the first studies to provide experimental evidence supporting the consolidation hypothesis. Electroconvulsive shock treatments impaired memory when administered to rats immediately after training. These findings were replicated and extended in extensive research with rats and mice in experiments using many kinds of treatments that disrupt brain functioning (McGaugh and Herz, 1972; McGaugh, 2000). The common finding was that the treatments affected memory when administered shortly after training and were less effective when administered several hours or longer after training. These early findings of retrograde amnesia induced by disrupting brain functioning after learning suggested the possibility that mild stimulation of the brain shortly after an experience might enhance memory (McGaugh, 1966). The finding of many studies that memory is enhanced by administration of low doses of CNS stimulants to rats and mice shortly after training but not after a delay provided strong support for this implication (Breen and McGaugh, 1961; Westbrook and McGaugh, 1964; McGaugh and Petrinovich, 1965; McGaugh, 1966, 1968, 1973; Krivanek and McGaugh, 1968; McGaugh and Roozendaal, 2009). Also, importantly, comparable

findings were obtained in studies using human subjects (Soetens et al., 1993).

ENDOGENOUS MODULATION OF MEMORY CONSOLIDATION

The findings of experimentally induced retrograde amnesia and memory enhancement also suggest a hypothesis that might provide an answer to the question of why, as Francis Bacon asserted, memory is assisted by passion. Emotional arousal induces the release of the adrenal stress hormones epinephrine and cortisol (corticosterone in rats). Thus, the levels of the hormones activated by arousing training experiences are increased while memories are undergoing consolidation. Gerard (1961, p. 30) noted that “as epinephrine is released in vivid emotional experiences, such an intense adventure should be highly memorable.” Much subsequent evidence supports this suggestion. As found with stimulant drugs, posttraining administration of epinephrine as well as corticosterone enhances memory for many kinds of training experiences (Micheau et al., 1984; Sandi and Rose, 1994; Zorawski and Killcross, 2002; Berlau and McGaugh, 2006; Roozendaal et al., 2006a). Furthermore, adrenoreceptor antagonists (e.g., propranolol) and glucococorticoid receptor antagonists block the effects of emotional arousal and adrenal stress hormones on memory consolidation (Gold and Van Buskirk, 1975; Gold and McGaugh, 1977; McGaugh, 1983; McGaugh and Gold, 1989; Roozendaal, 2000; Krugers et al., 2011; Roozendaal and McGaugh, 2011; Parfitt et al., 2012). Although most experiments investigating stress hormone influences have used memory of stressful training, such as stress induced by mild footshocks, posttraining administration of stress hormones enhances memory for many kinds of less stressful experiences, including memory for rewards (Dornelles et al., 2007).

AMYGDALA ACTIVATION AND MEMORY MODULATION

Thus, the experimental evidence provides strong support for the hypothesis that adrenal stress hormones enhance the consolidation of memory of experiences that induce their release. Additionally, the findings provide an initial step to providing an answer to William James’ question of why some memories endure. A next essential step requires understanding of how adrenal stress hormones act to influence brain processes involved in memory consolidation. When released into the blood, epinephrine passes poorly, if at all, into the brain (Arai et al., 1981). Considerable evidence indicates that epinephrine influences on brain function are mediated by activating adrenoreceptors located on the ascending vagus nerve that projects to brainstem nuclei (to the locus coeruleus

through the nucleus of the solitary tract) responsible for noradrenergic activation of other brain regions (Miyashita and Williams, 2006; McIntyre et al., 2012). Moreover, direct electrical stimulation of the ascending vagus after learning enhances memory in human subjects as well as rats (Clark et al., 1998, 1999; Clayton and Williams, 2000; Hassert et al., 2004). Cortisol passes freely into the brain, where it can activate glucocorticoid receptors throughout the brain.

Several findings suggested the amygdala, a collection of nuclei located in the medial temporal lobe, as a possible critical brain region involved in mediating stress hormone influences on memory consolidation. Findings of several early (Gold et al., 1975; McGaugh and Gold, 1976) as well as more recent studies (Bergado et al., 2006; Bass et al., 2012) indicated that, in rats, memory is enhanced by brief low-intensity posttraining electrical stimulation of the amygdala. Other early findings indicated that, in rats, b-adrenoreceptor antagonists infused into the amygdala after training impaired memory consolidation and that concurrent infusions of norepinephrine blocked the impairment (Gallagher et al., 1981; Kesner and Ellis, 1983). Other studies reported that systemically administered epinephrine induces the release of norepinephrine in the brain (Gold and van Buskirk, 1978) and that epinephrine enhancement of memory consolidation is blocked by intra-amygdala infusions of propranolol (Liang et al., 1986). There is now substantial evidence that norepinephrine or other noradrenergic agonists administered into the amygdala or selectively into the basolateral region of the amygdala (BLA) after training enhance memory for many kinds of training experiences (Hatfield and McGaugh, 1999; LaLumiere et al., 2003; McGaugh, 2004; Huff et al., 2005; Roozendaal et al., 2008). Also, posttraining intra-amygdala infusions of b-adrenoreceptor antagonists impair memory and block the memory-enhancing effects of both corticosterone and epinephrine administered systemically (Liang et al., 1986; Quirarte et al., 1997; Roozendaal et al., 2002, 2006b). Such findings strongly suggest that glucocorticoid-induced enhancement of memory consolidation requires noradrenergic activation of the amygdala. Noradrenergic activation induced by emotional arousal seems to enable glucocorticoid modulation of memory consolidation (Okuda et al., 2004).

The extensive evidence that memory is influenced by noradrenergic agonists and antagonists infused into the amygdala after training suggests that emotionally arousing training experiences should increase norepinephrine release within the amygdala. The findings of experiments using microdialysis and HPLC to assess norepinephrine release provide strong support for this implication. Footshock training increases the release of norepinephrine within the amygdala (Galvez et al., 1996; Quirarte et al., 1998), and rats that have greater increases in release subsequently display better retention (McIntyre et al., 2002). Additionally, several drugs that

enhance memory consolidation, including GABAergic and opioid peptidergic antagonists, increase the release of norepinephrine in the amygdala (Quirarte et al., 1998; Hatfield et al., 1999).

AMYGDALA INFLUENCES ON OTHER BRAIN SYSTEMS

Decades before initiation of the research discussed above investigating the involvement of the amygdala in memory consolidation, Gerard (1961, p. 30) noted that, “[because] the amygdala [acts] directly on cortical neurons to alter . . . their responsiveness to the discrete impulses that reach the cortex . . . these deep nuclei could easily modify the ease and completeness of experience fixation.” The amygdala is richly interconnected with other brain regions, including the cortex, known to be involved in processing different aspects of memory. Additionally, there is now considerable evidence supporting the prescient suggestion by Gerard (1961) that the amygdala influences memory consolidation through projections to other brain regions (McGaugh, 2000, 2002, 2004; Paré et al., 2002; Pelletier et al., 2005; Popescu et al., 2009; Stefanik et al., 2013). However, the interactions are not restricted to the cortex.

The findings of many studies using rats indicate that the hippocampus is involved in spatial learning (O’Keefe and Dostrovsky, 1971; Olton et al., 1979; Eichenbaum et al., 1990), whereas the caudate nucleus is involved in the learning of specific cues associated with responses (Packard and White, 1991; Packard and McGaugh, 1996; Packard and Goodman, 2012). Packard et al. (1994) found that posttraining activation of the amygdala (using microinfusions of d-amphetamine) enhanced memory for both place learning and cued response learning in a water maze. In contrast, hippocampal infusions selectively enhanced spatial memory, and caudate infusions selectively enhanced cued response memory. Posttraining, intra-BLA drug infusions enhance rats’ memory of a context as well as the memory of a brief footshock subsequently received in that context (Malin and McGaugh, 2006). Additionally and importantly, McIntyre et al. (2005) found that, in rats, noradrenergic activation of the BLA that enhanced memory consolidation increased the expression of activity-regulated cytoskeletal (Arc) protein in the hippocampus. Furthermore, posttraining inactivation of the BLA impairs memory consolidation and decreases hippocampal Arc protein expression. These findings are of interest in view of evidence indicating that Arc is involved in regulating synaptic plasticity and memory consolidation (Guzowski et al., 2000). Additionally, electrical stimulation of the BLA enhances the development of hippocampal plasticity as assessed by induction of long-term potentiation (Ikegaya et al., 1995; Akirav and Richter-Levin, 1999).

As noted by Gerard (1961), the amygdala also projects to the cortex. Electrical stimulation of the BLA activates the cortex, which is indicated by EEG desynchronization (Dringenberg and Vanderwolf, 1996; Dringenberg et al., 2001), and enhances cortical long-term potentiation (Dringenberg et al., 2004). Additionally, electrical stimulation of the amygdala enhances the development of plasticity in the auditory cortex (Chavez et al., 2012). It is well established that pairing of a tone stimulus with a reinforcing stimulus (e.g., footshock) alters the representation of the tone in the auditory cortex (Weinberger, 2004, 2007). The frequency-receptive fields shift to the frequency of the tone stimulus, inducing an increased representation of significant sounds. Pairing of a tone with BLA stimulation, which is neither rewarding nor punishing, induces a shift of the auditory tuning curve to that tone of the conditioning tone frequency. Moreover, the tuning curve continues to shift to the conditioning tone frequency over a period of 45 min after the training (Chavez et al., 2009). Importantly, the learning-induced shifts in responsiveness of the auditory cortex are maintained for several weeks (Chavez et al., 2013).

EMOTIONAL AROUSAL, ADRENAL STRESS HORMONES, AND HUMAN MEMORY

The findings of studies of the influence of arousal on human memory are consistent with the findings of studies using animal subjects: emotional arousal during or after learning enhances long-term memory, and the modulation involves epinephrine and cortisol. However, experiences do not have to be intensely emotional to influence memory strength. Many studies have reported that subjects presented with pictures or words judged to be only mildly emotional, whether positive or negative in effect, subsequently have stronger memories of those stimuli than pictures and words judged not to be emotional (Anderson et al., 2006; Liu et al., 2008; Henckens et al., 2009; Nielson and Lorber, 2009; Kensinger et al., 2011). Viewing of emotionally arousing pictures also enhances memory of a cognitive skill (Steidl et al., 2011).

In support of the view that emotional arousal modulates memory consolidation, several studies have reported that inducing arousal after subjects learn material enhances memory tested after a retention interval of 1 day or longer (Nielson et al., 1996; Anderson et al., 2006; Liu et al., 2008; Nielson and Lorber, 2009; Kensinger et al., 2011). In one study, subjects learned a word list and then watched an emotionally arousing pleasant (comedy) or unpleasant (surgery) brief video either immediately or after delays of up to 45 min. When viewed within 30 min, both the pleasant and unpleasant postlearning videos enhanced memory as assessed 1 week later (Nielson and Powless, 2007). Furthermore, the effect of postlearning

arousal is not restricted to laboratory experiments. College students who watched an arousing video clip after a lecture compared with students who did not watch the clip performed significantly better on a midterm examination 2 weeks later (Nielson and Arentsen, 2012).

There is also extensive evidence that arousal influences on memory consolidation involve both epinephrine and cortisol. Administration of the adrenoreceptor antagonist propranolol before subjects’ viewing of a series of pictures accompanied by an emotionally arousing story blocked the enhancing effects of emotional arousal on memory assessed at 1 week later (Cahill et al., 1994). Administration of epinephrine or cold pressor stress (induced by holding an arm in ice water), which induces the release of epinephrine and cortisol, immediately after presentation of emotionally arousing pictures enhances subjects’ memory of the pictures (Cahill and Alkire, 2003; Cahill et al., 2003). Furthermore, Hupbach and Fieman (2012) reported that arousal induced by exposure to cold pressor stress after a memory retrieval test increased salivary cortisol and enhanced memory of the test material when tested several days later.

Other studies have reported evidence that adrenergic activation selectively influences memory for emotionally arousing stimulation (Maheu et al., 2004). Cold pressor stress induced after listening to neutral and emotional words selectively enhanced memory of the emotional words on a test the next day. Furthermore, levels of cortisol and salivary α-amylase, a biomarker for noradrenergic activity, assessed immediately after the cold pressor stress correlated highly with subsequent memory performance (Smeets et al., 2008). Additionally, Segal and Cahill (2009) found that levels of salivary α-amylase assessed shortly after subjects viewed a series of emotional and neutral pictures correlated significantly and selectively with memory of the emotional pictures on a 1-week retention test (Segal and Cahill, 2009). Salivary α-amylase measured after exposure to emotionally arousing pictures also correlated highly with subsequent memory assessed by successful discrimination of pictures seen from other similar pictures (i.e., pattern separation) (Segal et al., 2012). Such discrimination is known to involve the hippocampus (Yassa and Stark, 2011).

Findings of human studies provide additional evidence that emotional arousal influences on memory involve activation of the amygdala. In an initial study using PET imaging, Cahill et al. (1996) found that amygdala activation induced by watching emotionally arousing films correlated highly with memory of the films as tested 3 weeks later. Subsequent studies using PET imaging reported similar findings (Hamann et al., 1999, 2002). Furthermore, studies using functional MRI imaging found that the relationship between amygdala activity during learning and subsequent memory varied directly with the intensity of emotional arousal and that

the valence—positive or negative—is not critical (Canli et al., 2000, 2002; Kensinger and Corkin, 2004).

Imaging studies have also provided evidence, consistent with evidence obtained with animal studies, that emotional arousal influences on consolidation of long-term memory involve interactions of the amygdala with other brain regions, including the hippocampus, during learning (Kilpatrick and Cahill, 2003; Dolcos et al., 2004; Kensinger et al., 2006; LaBar and Cabeza, 2006; Ritchey et al., 2008, 2011; Schwarze et al., 2012). Findings of human brain imaging studies using functional MRI provide additional evidence that emotional arousal influences on memory involve noradrenergic activation of the amygdala. Propranolol blocks amygdala activation induced by emotionally arousing stimuli as well as subsequent memory of the stimuli (Strange and Dolan, 2004; van Stegeren et al., 2005, 2006, 2007). Furthermore, administration of either the adrenergic drug yohimbine or hydrocortisone enhanced amygdala and hippocampus activation as well as memory as tested 1 week later (van Stegeren et al., 2010).

The findings of several studies suggest that intense or excessive activation of this noradrenergic system may contribute to the development of posttraumatic stress disorder (PTSD) (Pitman, 2006). Propranolol administered to traumatized patients within several hours after a traumatic experience expressed fewer physiological signs of PTSD when tested 1 month later (Pitman et al., 2002; Vaiva et al., 2003). Additionally, a study of the incidence of PTSD in wounded military personnel reported that patients given morphine within hours after the injury expressed fewer signs of PTSD when examined months after the experience (Holbrook et al., 2010). Because opiates inhibit the release of norepinephrine (McGaugh et al., 1988), a morphine-induced reduction in noradrenergic activation shortly after the trauma may have attenuated the development of PTSD.

EXCEPTIONAL HUMAN MEMORY

The findings summarized above provide the beginnings of understanding why, as William James wondered, some memories endure. Moreover, he suggested that the fact that many, perhaps most, memories are fleeting is adaptive. There is usually no need for memory of every detail of our daily experiences. As James (1890, p. 680) commented, “[s]election is the very keel on which our mental ship is built. If we remembered everything, we should, on most occasions be as ill off as if we remembered nothing.” The fictional character in Borges’ (1944) short story, “Funes the Memorious,” illustrated James’ observation. After he was thrown from a horse, Funes expressed an extraordinary ability to learn and remember. He “remembered not only every leaf of every tree of every wood, but also every one of the times that he had perceived . . . it” (Borges, 1944). Also,

he claimed to have “more memories than all mankind has had” (Borges, 1944). However, he also admitted, as James had anticipated, that his memory was like a garbage heap. Borges thus agreed with William James in stressing the importance of forgetting.

Luria (1968) subsequently documented the now well-known case of a subject referred to as S, who had an extraordinarily strong memory ability resembling the ability of Funes. Luria concluded that S’s memory capacity and durability were unlimited. Also, he asked, “[h]ow had he come by this capacity for indelible memory traces?” (Luria, 1968, p. 61). Although another individual was subsequently determined to have comparable memory ability (Hunt and Love, 1972), Luria’s question remains unanswered. It should also be noted that S’s phenomenal memory seemed to be of little help to him in his daily personal life.

A small percentage of autistic individuals are capable of highly exceptional but restricted memory abilities. Calendar calculation is one of the most commonly reported abilities. Some autistic individuals can readily state the day of the week for any specified date over a range of centuries, despite an inability to remember how to do simple addition and subtraction (Hurst and Mulhall, 1988; Heavey et al., 1999). Such complex memory-based ability, like the ability of subject S, remains unexplained.

Recent research has identified a few human subjects who have a remarkable memory ability referred to as highly superior autobiographical memory (HSAM) (Parker et al., 2006; LePort et al., 2012). The first subject identified to have this kind of memory (originally referred to as AJ and now known as Jill Price) wrote: “I am thirty-four years old and . . . have had this unbelievable ability to recall my past. . . . I can take a date, between 1974 and today, and tell you what day it falls on, what I was doing that day and if anything of great importance occurred on that day I can describe that to you as well” (Parker et al., 2006). Extensive testing confirmed her claims. She was remarkably accurate in recalling the experiences of most of the days of her life beginning at about the age of 11 years. Her extensive diary entries confirmed her memories of events that occurred on specific days. Her memory of significant public events is equally superior. Also, importantly, she does not do calendar calculation: unlike autistic savants, she cannot readily provide the day for dates when she was very young or future dates. After the publication of the paper by Parker et al. (2006), testing of many dozens of subjects who claimed to have strong autobiographical memory yielded several dozen subjects who surpassed age- and sex-matched controls in remembering the days and dates of personal and public events as well as details for each event (LePort et al., 2012). All responses were verified by checking available personal and public records. In contrast and perhaps surprisingly, HSAM subjects did not generally excel in learning and remembering information

as assessed by laboratory tests (e.g., learning pairs of words and series of digits). Their exceptional memory ability seems to be restricted to experiences of daily life as reflected in episodic remembering. Their memory is not like the memories of Borges’ character Funes the Memorious, Luria’s subject S, or autistic savants. They are also not like the memory experts, who have learned specific mnemonic tricks enabling the learning of specific kinds of information (Wilding and Valentine, 1997).

MRI scans revealed that several brain regions of HSAM subjects differed from those regions of controls. Several brain regions differed in size and shape (e.g., putamen and caudate) as well as coherence of fiber tracts (e.g., uncinate faciculus) as assessed by diffusion tensor imaging. These results are, of course, only correlational and do not provide critical evidence that these anatomical differences are the bases of or contribute in some way to HSAM. However, it is worth noting that several of the brain regions found to be structurally different in HSAM and control subjects have been implicated in previous studies of autobiographical memory (Kapur et al., 1992; Levine et al., 1998; Steinvorth et al., 2006).

UNUSUAL MEMORY: FLEETING AND LASTING

Studies of unusual memory have significantly influenced memory research as well as our understanding of the neural systems underlying memory. The clinical findings of Ribot (1882) that brain damage impairs most recent memories, sparing older ones, were the first to reveal that lasting memories are consolidated slowly over time. The findings of seminal studies of the patient HM (Scoville and Milner, 1957; Milner, 1972) forced the novel conclusion that different forms of memory are enabled by different brain systems (Mishkin et al., 1984; Packard and McGaugh, 1996; White and McDonald, 2002), and thus they significantly altered research investigating brain systems and memory. The novel finding that some human subjects create highly lasting memories of episodes of their daily experiences as well as memories of significant public events may, ultimately, lead to findings that provide new understanding of how our brains retain and retrieve memories.

Studies have not, as yet, investigated whether the strong memory of HSAM subjects involves experience-induced activation of stress hormones and activation of the amygdala. It may be that the modulatory systems of HSAM subjects are more highly activated by experiences or more sensitive to modestly arousing experiences.

However, HSAM subjects do not remember in precise detail all of their experiences. Like the rest of us, they remember best the more significant events of daily life. Their memories are not like the memories of Funes the Memorious. However, HSAM subjects differ from the rest of us in that

they can retain their episodic memories for decades. If lasting memory is important for survival, why is it that so few individuals have this kind of long-lasting memory? It might be that these subjects’ memory systems are genetically programmed to retain acquired information. Although the evidence, to date, indicates that none of the HSAM subjects have relatives who have strong memory, additional research is needed to determine whether this ability may have a genetic basis.

We might also wonder whether this ability might have been more common and more commonly used in centuries past. After all, the inventions of the printing press, computers, and cell phones in recent centuries have made it less necessary for us to create lasting records of our experiences. It is said that, before writing was available to keep records of important events, such as a wedding or granting of land, a child was selected to observe an event and then thrown into a river so that the child would subsequently have a lifelong memory of the event. As noted above, for most of us, “[o]ur brains, remarkable as they are, [can]not begin to contain and give equal weight to our every moment of life” (Glore, 1987). Selectively remembering our more important experiences seems to be the best strategy. It is what we generally do, thanks to the modulating influences of emotional arousal on lasting memory.

ACKNOWLEDGMENTS

Research was supported by National Institute of Mental Health Public Health Service Grant MH12526 and the Gerard Family Trust.