Modulation of long-term memory by arousal in alexithymia: The role of interpretation

Abstract

Moderate physiological or emotional arousal induced after learning modulates memory consolidation, helping to distinguish important memories from trivial ones. Yet, the contribution of subjective awareness or interpretation of arousal to this effect is uncertain. Alexithymia, which is an inability to describe or identify one’s emotional and arousal states even though physiological responses to arousal are intact, provides a tool to evaluate the role of arousal interpretation. Participants scoring high and low on alexithymia (N = 30 each) learned a list of 30 words, followed by immediate recall. Participants then saw either an arousing (oral surgery) or neutral video (tooth brushing). Memory was tested 24-h later. Physiological response to arousal was comparable between groups, but subjective response to arousal was impaired in high alexithymia. Yet, delayed word recognition was enhanced by arousal regardless of alexithymia status. Thus, subjective response to arousal, i.e., cognitive appraisal, was not necessary for memory modulation to occur.

Introduction

Alexithymia is a construct describing a cluster of characteristics that includes difficulty identifying, describing, and communicating feelings; difficulty distinguishing between feelings and bodily sensations; constricted imaginal processes; and an externally oriented cognitive style (Sifneos, 1973, Taylor, 2000, Taylor et al., 1991). Alexithymia is also a relatively stable trait under changing degrees of emotion, stress, and pathology (Luminet et al., 2001, Mikolajczak and Luminet, 2006, Subic-Wrana et al., 2005). Although its precise etiology is unknown, alexithymia has been postulated to be associated with emotional dysregulation (Taylor, 1994). As such, individuals with alexithymia lack awareness (i.e., subjective acknowledgment) of their emotional and arousal states and they tend to suffer from psychiatric and psychosomatic disorders at a high rate (e.g., Lumley et al., 1996, Taylor and Taylor, 1997).

The link between alexithymia and disease has been postulated to be due to prolonged exposure to adrenal hormones as a result of unacknowledged and unresolved stress and arousal (Chrousos, 1998, Martin and Pihl, 1985, Papciak et al., 1985). This postulation has received much support, in spite of differences across study methods and specific results (e.g., Friedlander et al., 1997, Martin and Pihl, 1986, Stone and Nielson, 2001, Wehmer et al., 1995). For example, our experimental study showed that high alexithymia was associated with heightened baseline arousal level and intact physiological response to an arousing video clip of live-action oral surgery (Stone & Nielson, 2001). However, subjective response to the clip was impaired in high alexithymia. Importantly, the distinction between subjective and physiological response to arousal in alexithymia may provide a valuable tool for investigating the role of emotion and arousal in memory consolidation.

Emotional and arousing events are recollected with greater frequency than similar but neutral events. This is likely an adaptive function, effectively highlighting important events toward protecting and preparing an organism for similar future events (McGaugh, 2000). Memory consolidation, believed to underlie this highlighting process, is the outcome of a complex set of neurobiological processes occurring after the initial formation of a memory (McGaugh, 2000, Müller and Pilzecker, 1900, Nielson and Powless, 2007, Revelle and Loftus, 1992, Torras-Garcia et al., 1997).

A variety of substances, including glucose and the adrenal hormones epinephrine and norepinephrine, are released into the bloodstream during times of arousal, stress and emotion, which indirectly alter brain stem and amygdalar function (Gold, 2005, Hamann, 2001, McGaugh, 2000, McGaugh et al., 1996). When administered soon after learning, these substances alter memory in a classic ‘‘inverted-U’’ manner, whereby moderate amounts enhance performance and low or high amounts can impair or fail to influence performance (McGaugh, 2000, Yerkes and Dodson, 1908). Although emotional events naturally involve arousal onset during the event itself, arousal typically also persists for some time afterward. Therefore, like other arousal sources, emotion can have physiological effects on memory consolidation, as well as on encoding and attention. Indeed, although arousal can facilitate detection and encoding for long-term retention, it can also hinder retrieval in the short-run and not reveal its enhancing effects for hours or even days (Kleinsmith and Kaplan, 1963, Revelle and Loftus, 1992, Torras-Garcia et al., 1997, Walker, 1958). These effects are also time-dependent, such that doses administered during or soon after learning are effective, but those administered after long intervals are ineffective (e.g., Gold and van Buskirk, 1975, Nielson and Powless, 2007).

The preponderance of human studies examining the effects of arousal on learning or memory have used either inherently emotional materials or other interventions given before or coincident with the learning task. As such, it is impossible to decipher in these studies whether the arousal effect is on attention, encoding, consolidation, or some combination of these memory phases. However, several recent studies have demonstrated memory modulatory effects in human participants by comparable mechanisms of action as have been shown in rodent studies. These studies have spanned a variety of tasks and post-learning treatments, including norepinephrine (Southwick et al., 2002), glucose (Manning, Parsons, & Gold, 1992), nicotine (Colrain, Mangan, Pellett, & Bates, 1992), and non-invasive treatments such as muscle tension (Nielson and Jensen, 1994, Nielson et al., 1996), stress (Cahill, Gorski, & Le, 2003), and negative and positive emotional arousal (Nielson and Bryant, 2005, Nielson and Powless, 2007, Nielson et al., 2005).

Despite the extensive animal and growing human literatures on the biological foundations of memory modulation by arousal, a number of cognitive studies have not found significant effects of arousal manipulations on memory performance. Such null findings have led to the contention that arousal only influences memory if the arousal source is semantically related to the memoranda (e.g., Christianson and Mjörndal, 1985, Varner and Ellis, 1998) or if the memoranda are inherently arousing (Buchanan and Lovallo, 2001, Cahill et al., 2003, Liu et al., 2008). These studies have greatly contributed to the discussion of emotion and memory. However, such conclusions were premature due to several limitations of these studies (cf. Nielson et al., 2005), including the use of very short delays before testing, which do not allow for the process of memory consolidation to occur (Kleinsmith and Kaplan, 1963, McGaugh, 2000, Quevedo et al., 2003, Revelle and Loftus, 1992, Torras-Garcia et al., 1997, Walker, 1958).

Moreover, the semantic relatedness contention about arousal does not consider the growing number of studies demonstrating that arousal that is induced after learning modulates long-term memory for materials that have no semantic relation to the arousal stimulus (Colrain et al., 1992, Manning et al., 1992, Nielson and Bryant, 2005, Nielson and Jensen, 1994, Nielson and Powless, 2007, Nielson et al., 1996, Nielson et al., 2005, Southwick et al., 2002). Therefore, there would be a benefit to the literature from studies designed to dissociate the effects of physiological arousal response from the effects of arousal interpretation on memory. Alexithymia presents a unique opportunity to test this idea because of the inherent dissociation of arousal response from its interpretation. That is, studies of participants who endorse high and low alexithymia tendencies present the opportunity to evaluate whether memory modulation by arousal is influenced by impaired subjective, but not physiological, response to arousal.

Cognitive abilities have not been often empirically studied in alexithymia, and memory modulation has been entirely unstudied in alexithymia. However, in line with intact physiological responding in alexithymia, a recent visual evoked potential study demonstrated intact perceptual processing of emotional and neutral pictures by alexithymics (Franz, Schaefer, Schneider, Sitte, & Bachor, 2004). Yet, there are suggestions that the cognitive schemas necessary for processing the meaning of emotion are deficient in alexithymia (Taylor, Bagby, & Parker, 1997). Consistent with this notion, one study found increased Stroop interference in alexithymia for illness words relative to negative emotion words, but not relative to neutral words (Lundh & Simonsson-Sarnecki, 2002).

A few recent studies have also examined memory performance in alexithymia, with mixed results. For example, one study demonstrated that both positive and negative autobiographical memories were equally well recalled in high and low alexithymia (Lundh, Johnsson, Sundqvist, & Olsson, 2002). However, another study found no alexithymia differences in recall of neutral words or emotion words that were only characterized by a feeling of knowing (“know”), but high alexithymia participants recalled significantly fewer emotion words that could be recollected with contextual detail (“remember”, Luminet, Vermeulen, Demaret, Taylor, & Bagby, 2006). Finally, one study demonstrated that those with high alexithymia had reduced implicit memory for positive emotion words (Suslow, Kersting, & Arolt, 2003). Thus, although little study of memory in alexithymia has been yet undertaken, the available results suggest memory may be normal for neutral stimuli. Results are mixed regarding emotive stimuli. Yet, these studies used only immediate tests of memory retrieval; long-term retrieval in alexithymia remains unstudied.

Therefore, the purpose of the current study was to evaluate long-term memory in alexithymia with neutral stimuli. Specifically the purpose was to examine whether arousal-induced memory modulation occurs comparably in high and low alexithymia, toward determining whether subjective awareness of arousal state influences memory modulation. The design was based upon that used in our previous alexithymia study (Stone & Nielson, 2001), with the addition of a memory task. Briefly, prospective participants (>300) were screened using the Toronto Alexithymia Scale-20 (TAS-20; Bagby et al., 1994, Bagby et al., 1994). Those scoring in the top and bottom quintiles (n = 30 each) viewed a 30-noun word list, after which recall was tested. Afterward, either an arousing (oral surgery) or neutral (tooth brushing) video clip was presented (Nielson et al., 2005, Stone and Nielson, 2001), with quasi-random assignment to groups. Recognition was tested 24 h later, with scores corrected for guessing (see Supplementary materials for detailed methods). It was hypothesized that arousal after learning would enhance delayed retrieval (e.g., Nielson and Bryant, 2005, Nielson and Jensen, 1994, Nielson and Powless, 2007, Nielson et al., 1996, Nielson et al., 2005). Moreover, because high alexithymia was expected to be associated with intact physiological response to arousal despite impaired subjective response (Stone & Nielson, 2001), it was further hypothesized that alexithymia would not differentiate or interact with memory performance.