Emotional context modulates subsequent memory effect

Abstract

Emotions have been shown to modulate memory processes. However, the neuronal substrate underlying these modulatory effects is largely unknown. Using event-related functional magnetic resonance imaging (fMRI), we investigated whether the context of emotional encoding modulates brain activation predictive for subsequent recall of emotionally neutral material. While inferior frontal activation predicted recall in general, our data show that in a positive encoding context, recall was predicted by activation of right anterior parahippocampal and extrastriate visual brain areas, whereas in a negative encoding context, recall was predicted by activation of the amygdala. Thus, we could demonstrate that successful episodic encoding is differentially modulated by emotional context. These results contribute to the understanding of the interaction of emotion and cognition and moreover are of general relevance for studies of episodic memory.

Introduction

In recent years the importance of the comprehension of emotions has been convincingly demonstrated Borod 1992, Damasio 1994, LeDoux 1996, Rolls 1999. Emotions have been shown to play an important part in cognitive functions with the amygdala being involved in negative emotions and the orbitofrontal cortex and extrastriate visual areas in positive emotions Blood and Zatorre 2001, Critchley et al 2000, Dolan et al 1996, Lang et al 1998, Morris et al 2001, O’Doherty et al 2001. Considerable evidence shows that memory for emotional material is enhanced (Cahill et al., 1994), particularly for negative material with the amygdala crucially being involved Adolphs et al 1997, Cahill et al 1996. Meanwhile, the role of the amygdala in human episodic memory for emotional material has been extensively demonstrated with functional neuroimaging methods (for example, Canli et al 2000, Hamann et al 1999.

However, all of these studies scrutinized memory for emotional material. In other words, they investigated the emotional content of memory. This approach has the disadvantage that effects of different emotional states and effects of the material itself cannot be distinguished.

On the other hand, in behavioral psychology the influence of emotional context on memory of emotionally neutral material has been investigated in a series of studies Ashby et al 1999, Ashby et al 1999, Fiedler 1988, Fiedler 1991. It has been demonstrated that active recovering of information from episodic memory as it is the case in free recall is enhanced in positive emotional context during encoding compared to neutral or negative context. This backs up the hypothesis that positive and negative emotional context favor different information processing procedures, which in turn should be subserved by different neural substrates.

The neural correlates of the influence of emotional context on encoding have not yet been investigated in human functional neuroimaging. The influence of emotional context has been investigated recently only for retrieval. During retrieval of neutral words which had been presented in emotionally valenced sentences during encoding, enhanced activity was found during retrieval in amygdala and hippocampus for items from negative vs neutral contexts and in orbitofrontal cortex for items from positive vs neutral contexts (Maratos et al., 2001).

We wanted to study the modulatory influences of emotional context on encoding of emotionally neutral material. For this purpose we studied brain activation during encoding using fMRI and assessed the neural correlates of subsequent successful recall. A subsequent memory effect (SME), i.e., greater brain activation during encoding for later successfully remembered compared to forgotten material, has been demonstrated in a number of human studies by means of functional neuroimaging techniques and intracerebral recordings (Paller et al., 2002). It has been shown that activation in frontal and medial temporal regions during encoding predicts subsequent memory for verbal Otten et al 2001, Wagner et al 1998 and nonverbal (Brewer et al., 1998) material. The SME has been investigated until now irrespectively of emotional context by using mostly recognition tasks, whereas two recent studies used free recall Fell et al 2001, Strange et al 2002. For our study we used a free recall task instead of a recognition task with the rationale being that the influence of emotional context is found particularly in tasks requiring an active recovering of information from memory, which is more likely in recall than in recognition Doerksen and Shimamura 2001, Fiedler 1991.

Based on the above mentioned studies we predicted an activation in left inferior frontal cortex for the main effect of subsequent memory. Furthermore, we predicted behavioral memory enhancement for positive vs negative emotional encoding context when using a free recall task. Finally, we expected to find neural correlates for this behavioral effect either in regions known to be implicated in subsequent memory, i.e., prefrontal cortex, parahippocampal, or fusiform cortex or in regions known to be involved in emotional information processing, i.e., the amygdala and extrastriate regions.