Cognitive and affective theory of mind abilities in alcohol-dependent patients: The role of autobiographical memory

Abstract

Background

Many studies of patients with alcohol dependence (AD) have highlighted their difficulty in identifying both their own emotional state and those of a social partner. We examined (1) the cognitive and affective theory of mind (ToM) abilities of AD patients and (2) how the efficiency of their autobiographical memory (AM) can affect the effectiveness of ToM ability.

Method

In a cross-sectional design, AD patients (N = 50) and healthy controls (N = 30) completed a ToM movie paradigm (Versailles-Situational Intention Reading, V-SIR) in which they inferred the intentions of characters in movies depicting social interactions, and the “Reading the Mind in the Eyes” Test (RMET), which assessed the emotional dimension of the ToM. AM was investigated using the “Autobiographical Memory Interview” (AMI) to assess both episodic and semantic components of AM.

Results

Concerning ToM, patients with AD showed lower performance in the RMET than control participants, whereas no difference was observed on the V-SIR test. AD patients had lower scores than controls on the AMI, for both episodic and semantic components and for different periods of life. A multiple linear regression analysis also showed that AM deficits might predict lower ToM performance, especially for the RMET task.

Conclusions

Patients with AD have a specific affective ToM deficit. They used episodic memories to perceive the emotions of others, whereas controls used preferentially semantic memories to perform the task. Both these deficits could constitute a risk of relapse and should be a target for psychotherapeutic interventions.

Introduction

In recent years, many studies have highlighted social cognition impairments in alcohol-dependent (AD) individuals (Uekermann et al., 2007), with a particular emphasis on the treatment of emotional signals. These studies have employed paradigms that use facial emotion recognition (Philippot et al., 1999, Townshend and Duka, 2003, Kornreich et al., 2001, Foisy et al., 2007, Maurage et al., 2008), identification of discourse prosody (Monnot et al., 2001, Uekermann et al., 2005, Kornreich et al., 2013), identification of postures (Maurage et al., 2009), or the ability to differentiate self and other emotional states (Gandolphe et al., 2013, Bochand and Nandrino, 2010). Researchers have put forward the view that these emotional difficulties in substance-dependent patients may lead to them experiencing stress in social situations, which may then result in alcohol consumption and relapse (Kornreich et al., 2002, Kornreich et al., 2013, Uekermann et al., 2007). Although these previous studies have investigated the emotional state attribution competences of substance-dependent patients through basic emotion recognition tasks, a comparison with more cognitive aspects of social cognition is still lacking. As noted by Russell et al. (2009), the concept of “hot” and “cold” ToM, originally posited by Brothers and Ring (1992), has been reintroduced and renamed “affective” and “cognitive” ToM (Blair, 2005, Shamay-Tsoory et al., 2006). Accumulating neuroanatomic evidence supports this idea that the cognitive theory of mind (belief attribution) and affective ToM (empathy mechanisms) may be associated with overlapping, but distinct, neuronal networks (Völlm et al., 2006). In fact, while these two processes involve the medial prefrontal cortex, the temporoparietal junction, and temporal poles, empathy is associated with enhanced activation of the paracingulate, anterior and posterior cingulate and amygdala, and belief attribution accompanies increased activation in the lateral orbitofrontal cortex, middle frontal gyrus, cuneus and superior temporal gyrus.

To describe precisely the ongoing debate related to the nature of ToM (considering both affective and cognitive components; Russell et al., 2009), we proposed to understand the social cognitive deficits in alcohol-dependent individuals by assessing both the emotional and cognitive processes of the ToM. To assess the affective ToM processes, we referred to Baron-Cohen's (1995) model, which is intended to assess different levels of mental state attribution processes. The tools developed in this framework, specifically the Reading the Mind in the Eyes Test (RMET, Baron-Cohen et al., 2001), assess the first stage of mental state attribution (the “unconscious, rapid and automatic” decoding of the “language” of the eyes) without requiring cognitive processes to deduce content (Baron-Cohen et al., 2001, p. 241). Imaging data have confirmed that the RMET activates regions typically associated with emotion processing, specifically the amygdala, and an absence of activation in this region is seen in those with ToM difficulties (Shaw et al., 2005). Moreover, using the RMET, an impairment of theory of mind abilities has been shown in high-risk substance-abuse individuals (i.e., non-dependent young adults with a positive family history of alcohol dependence) (Hill et al., 2007), amphetamine abusers (Kim et al., 2011, Henry et al., 2009), and alcohol-dependent patients (Maurage et al., 2011a, Thoma et al., 2013). Maurage et al. (2011a) have confirmed that substance-dependent patients have impaired performance of theory of mind for positive, negative and complex emotions. These results support the idea of a specific impairment of emotional attribution, rather than an overall difficulty in mental state attribution in alcohol-dependent patients.

In addition, Uekermann et al. (2007) assessed humor processing (presenting joke stems with different alternative endings reflecting incongruity and/or resolution) as theory of mind had previously revealed affective humor-processing deficits in alcohol-dependent individuals, explained as a reduced mentalizing ability.

Nearly all of the studies assessing ToM ability in AD participants have used paradigms that measure the emotional dimension without directly assessing overall social cognitive abilities involving cognitive aspects of attribution intents. The study by Kornreich et al. (2002) showed that AD patients have interpersonal relationship difficulties (measured with a self-report questionnaire; IRI), which are associated with difficulty in accurately recognizing facial expressions. Moreover, Thoma et al. (2013) confirmed a deficit with regard to mental state decoding, and also underlined an impaired mental state reasoning measured with the Faux Pas Story Test.

These studies assessing social cognition abilities in AD patients have used self-report questionnaires and experimental tasks without assessing these competences directly in ecological situations. As proposed by Bazin et al. (2009), we investigated theory-of-mind skills using more ecological tasks, which offer a good trade-off between enriched social contexts (video scenes with sound) and precisely defined cognitive measures targeting aspects of social cognition. We used the V-SIR to assess the subject's ability to infer mental states through the presentation of short video clips showing complex interactions between one or more individuals (Bazin et al., 2009).

Thus, in the present study, we assumed a ToM disorder in AD individuals that particularly affects the emotional aspects of processing compared to the cognitive components. We hypothesized that individuals with AD would have a lower score on the emotional ToM task (RMET) in comparison to control participants, but an equivalent performance with control participants on cognitive and ecological ToM tasks.

Moreover, we sought to identify factors that might influence the processing of mental state attribution. Recent literature supports the idea that the effectiveness of autobiographical memory could impact ToM competences. This implies that an individual would use his or her own knowledge concerning cognitive ability to imitate a series of probable actions mentally (Davies and Stone, 1995). Such a mechanism supposes a remembering of the subject's previous actions to be projected to the mind of another person. In particular, semantic autobiographical information enhances coherence of self-knowledge and identity over time, and serves to guide social interactions. This assumption is supported by recent neurofunctional observations highlighting that the recall of personally experienced events (autobiographical memory) and inferring the mental states of others (mentalizing or theory-of-mind) share a neural substrate (Spreng and Mar, 2012). The meta-analysis by Spreng et al. (2009) identified an overlap between reliable patterns of activity associated with AM and ToM. This overlap included the medial prefrontal and posterior cingulate cortex, frontal pole, inferior frontal gyrus, medial temporal lobe (parahippocampus, amygdala), superior temporal sulcus and middle temporal gyrus, and the angular gyrus (Spreng et al., 2009; Spreng and Mar, 2012). This functional overlap of neural networks would facilitate the integration of personal and interpersonal information and allow the transformation of personal experiences into social conceptual knowledge (Spreng and Mar, 2012).

In addition to this functional overlap, episodic memory and ToM emerge close in time in ontogenetic development and episodic memory development is specifically linked to the growing ability to introspect an ongoing experience in children between 3 and 6 years old (Naito, 2003, Perner et al., 2007). Lastly, the relationship between episodic memory and ToM is also supported by specific impairments observed in different pathologies. For example, Adler et al. (2010) highlighted the relationship between autobiographical memory and ToM competences in participants with high-functioning autism and Asperger's syndrome using two mental state attribution tasks [the RMET and the strange story task (SST; Happe, 1994)], two memory tasks [a modified version of the Personality Trait Questionnaire, (PTQ; Klein et al., 1999)] and the Rey-Auditory Verbal Learning Test [RAVLT; (Spreen and Strauss, 1998)]. The authors observed a positive correlation between RMET scores and the performance of episodic memory (subscores of the PTQ).

Moreover, Corcoran and Frith (2003) have shown that, when assessed with tools such as the AMI, individuals with schizophrenia tend to recall only salient events related to negative experiences or strange feelings in response to autobiographical events (Kopelman et al., 1990). These patients with schizophrenia use specific past experiences to interpret the situation, the behavior and the intention of a partner during social interaction (Corcoran and Frith, 2003).

In the light of this research, we proposed to study how a deficit in autobiographical memory in AD patients affects the effectiveness of ToM. Impaired episodic memory is typically observed in alcoholic patients and concerns both the encoding and the retrieval without an apparent relationship with executive dysfunction (Pitel et al., 2007, Le Berre et al., 2010). When the temporal distribution of AMs recalled in response to cue words is examined, AD participants recall fewer memories from their recent past (last five years) and more memories from their early adulthood than healthy controls do (Fitzgerald and Shifley-Grove, 1999). D’Argembeau et al. (2006) have shown that non-alcoholic amnesic patients report more general memories of specific events. However, when a specific memory is recalled, it is with equal phenomenological detail as in control participants. Whiteley et al. (2009) confirmed this lower specificity of autobiographical memories in AD subjects, which can be explained by the impairment of the frontal lobe associated with the consumption of alcohol, causing the failure of strategic processes involved in retrieving memories.

In the present study, we hypothesized that the performance would be impaired in the AD patients, in comparison with control participants, for the recall of both episodic and semantic autobiographical memories. In addition, we expected that scores on ToM tasks (RMET and V-SIR) would be determined by performance in episodic and semantic autobiographical memory.