Effect of Negative Memory Bias and Autobiographic Memory Specificity on Social Decision Making and Depression

The Effect of Negative Memory Bias and Autobiographic Memory Specificity on the Relationship Between Social Decision Making and Depression

Social decisions are critical for day-to-day functioning, building interpersonal relationships and determining the quality of social life (Rohde, 2016). Social decisions, affecting oneself and others, can range from simple choices of deciding what to wear for work to more complex decisions such as resolving conflicts in relationships. The ability to make effective social decisions is impaired when social functioning is disturbed, such as in depression (Wang et al., 2014

Hinterbuchinger et al., 2018). Alongside deficits in making social decisions, indivdiuals with depressive symptoms or major depressive disorder (MDD) present strong biases in their autobiographical memory. Past research indicates that we inform decisions at hand by drawing on memories of past choices (Bornstein, Khaw, Shohamy, & Daw, 2017). In doing so, bias in recalling autobiographic memories impacts decisions involved in social scenarios. Notably, studies reveal that depressed indivdiuals have more negative memories and lack autobiographic memory specificity, or remember generalised memoriesof events that lack personal, specific details of time and place (AMS; Matsumoto & Mochizuki, 2018). The literature has yet to identify these distinct pattern of memory biases as regulators of social decision making in depressed individuals, which would posit pertinent implications for interventions (Kuyen & Dalgleish, 2011; Hitchcock et al., 2018). Thus, the present study aims to investigate whether AMS or the bias toward negative memories moderates the relationship between depressive symptoms and social decision making skills.

Literature in memory has established a strong association between decision making and our memories of past events that heavily influence and guide everyday decisions (Bornstein et al., 2017). Models in autobiographic memory retrieval paths demonstrate how modulations to episodic memory can easily change our decision making skills (Conway & Pleydell-Pearce, 2000). Our event-specific memories are either activated by a cue and automatically retrieved, or intentionally searched for relevance or similarity to the imminent decision. Presence of cognitive bias along the second path has implications for our memories that could predict our decisions (Conway & Pleydell-Pearce, 2000). For instance, a person with a biased attention to social rejection might remember the times they were rejected from entering a club more than the times they were invited to other parties and choose to shy away from future invitations. This is consistent to Bornstein et al.’s (2017) findings where they manipulated bias in participants’ episodic memory by reminding outcomes from their past decisions. The study used neutral memory stimuli as reminder cues for each previous outcome, which increased the likelihood of particular memories being retrieved and enabled participants’ choices to be predicted. Findings from reward-based decision making tasks such as that of Murty, FeldmanHall, Hunter, Phelps, & Davachi (2016) also indicate that episodic memory was used to make adaptive choices. Participants formed stronger associative memory for high-reward cues and reengaged more with stimuli associated with high-rewards.

Associations formed by episodic memory further could generalise to complex choice behaviours, including social decision making (Murty et al., 2016). Duncan and Shohamy (2016) explored this by distinguishing how participants used memories of past events for decision making in a context-concordant manner. That is, indivdiuals were most likely to retrieve episodic memories to make their choice in familiar situations. When faced with a novel circumstance, memories were drawn from similar past choices to make a generalised, optimal decision. Further, experiencing these novel situations expanded participants’ memories of choice environments to aid future decisions (Duncan & Shohamy, 2016). The retrieval process of accessing and using the episodic memory bank for making decisions not only influences simple stimuli-based decision making but also complex problem solving undeniably within engagements in psychosocial contexts. As yet however, we have limited understanding of the direct impact of autobiographic memories on social decision making.

The autobiographic memories driving our daily decisions is susceptible to bias and this is particularly problematic for indivdiuals with depression, who show distinct patterns of disturbance in memory recall and maintenance (Dalgleish & Werner-Seidler, 2014). The depression-memory literature reveals that depressed indivdiuals have a systematic bias toward negative memories (Gotlib & Joorman, 2010). This bias can be observed in studies involving memory retrieval such as facial recognition tasks (Ridout et al., 2003) and free-recall task of emotive words (Gotlib & Joormann, 2010). Further, individuals with MDD exhibit hypersensitive neural responses to social rejection or negative cues (Kumar et al., 2017). Kumar et al. (2017) found that this hypersensitivity impairs social functioning and decision making skills by instigatingthe need to cope with aversive social signals, and consequently symptoms of interpersonal stress and withdrawal from social situations. In explanation, the Social Risk Hypothesis (Allen & Badcock, 2003) states that depression developed as an adaptive mechanism against social risks and negative outcomes such as rejection. Frameworks such as the mood congruency bias also suggest that stimuli that better reflect individuals’ current emotional states are more memorable. Perhaps as functions of these mechanisms, indivdiuals with depression show increased sensitivity and biased attention to negative memories and stimuli indicating social risk and

Depressed indivdiuals, particularly adolescents, are also biased toward vague memory and have reduced access to specific details of past memories. The AMS literature shows strong evidence of bias in at-risk and MDD-diagnosed adolescents toward categoric negative cue words and over generalised memories (Park, Goodyer, & Teasdale, 2002; Kuyken & Dalgleish, 2011). Conversely, Askelund, Schweizer, Goodyer and Harmelen (2019) recently reported that positive and specific memory style related to lower depressive symptoms and less negative perceptions of the self over the course of a year for at-risk adolescents with early life stress. Thus, current literature reveals negative, over generalised memory as a vulnerability marker and positive memory specificity as a resilience factor for depression in adolescents (Gutenbrunner, Salmon, & Jose, 2018). Identifying the above patterns in memory presentation has led to the implementation of memory training interventions focusing on improving and encouraging more positive and specific memories. 

While the direct effect of these memory biases on the association between depression and social decision making is uncertain, training interventions that directly target the memory biases show promising outcomes in clinical populations. Thus far, Memory Specificity Training (MEST) and positive memory elaboration in randomized controlled trial for adolescents has observed reductions in key depressive traits such as cognitive avoidance, rumination and improvement in problem solving skills (Neshat-Doost et al., 2013; Dalgleish & Werner-Seidler, 2014). Also, clinical trials of cognitive behavioural therapy for depressed patients showed that over general memory can be targeted through brief treatments (McBride, Segal, Kennedy, & Gemar, 2007). Despite these known effects of targeting negative memories and lack of memory specificity in such interventions, the direct moderating effect of AMS on the association between depressive symptoms and social decision making remains unclear.

The current study aims to investigate whether AMS moderates the relationship between performance on a social decision making task across levels of depressive symptoms. We also aim to observe whether the extent of bias toward negative memories reflect levels of depressive symptoms. To do so, we will recruit and compare the performance of low, moderate and high depression symptom groups on a social decision making task (SDMT). The current study will employ a modified version of the decision making task developed by Bornstein et al. (2017), which will instead include a social decision component.

In light of the above literature, we hypothesise that:

1. Participants with high depressive symptoms will show more bias toward negative (rejection) memory cues compared with positive (acceptance) memory cues at the recall test of the social decision making task.
2. The degree of this bias toward negative outcome cues in the social decision making task will reflect levels of depressive symptoms. That is, the discrepancy in choice behaviour to choose optimal (correct) outcome compared with non-optimal (incorrect) outcome on recall trials following negative or positive outcome memory cues will be most pronounced in participants with high depressive symptoms, followed by those with moderate, then low symptoms.
3. The association between the levels of depressive symptoms and performance in social decision

 making will vary as a function of participants’ AMS, where the effect will be stronger for those with over general memory than those with more specific autobiographical memory.

Method

Design

 The current study is an experimental study that compares social decision making performance of individuals who have high versus moderate versus low depressive symptoms. We employ a 3 x (2) mixed design where the between-subject factors are levels of participant depressive symptoms (low, moderate and high) and within-subject factors are outcome types (positive vs. negative) on the SDMT. The dependent variables are percentage of optimal choice made on the SDMT and percentage of specific memories on the AMT.

Participants

  A total of 100 participants aged 17-30 were recruited from the University of New South Wales psychology first-year student cohort, who were allocated into low and high depressive symptom groups through pre-screening questionnaires. Participants received course credit incentives for their contribution.

Materials & Stimuli

 The present study will use a short version (13 questions) of the Mood and Feelings Questionnaire (sMFQ) to measure participants’ symptoms of depression. The Strengths and Difficulties Questionnaire (SDQ) will also be completed for self-report of internalising and externalising problems and prosocial behaviour. Participants will also be asked to take a written AMT as a measure of their AMS, where they will be asked to write about a specific memory as a response to 10 cue words.

For the social decision making task, or the modified version of the decision task by Bornstein et al. (2017; see Experiment 2, bandit task), participants will be asked to imagine that they are to organise a large social event with as many attendees as possible. In each trial, they are to imagine going around the university campus to recruit attendees and to knock on either blue or yellow doors that could lead to either a positive (person behind the door attends the party) or a negative (person behind the door rejects party invitation) outcome. After choosing a door, participants will be shown unrelated neutral object stimuli as memory cue ‘posters’ on the door (such as a kettle), which is then followed by either a smiling face indicating an attendee, or a neutral face, indicating that the person will not be attending the event (see Figure 1).

 (A)            (B)                 (C)                 (D)

Figure 1. Sample trial showing stimuli sequence. (A) shows the initial door choice, followed by the neural stimuli shown in (B). (C) and (D) show the positive or acceptance and negative or rejection outcomes correspondingly.

The SDMT has a total of 162 trials consisting of 130 social decision making trials followed by 32 memory recall trials. At the recall trials, which will occur sometimes during the SDMT for practice and be measured at a final trial after the task, participants will be asked to recall the presence of an object memory cue and whether participants living behind the corresponding door were coming to the party or not. Lastly, a facial recall task will follow after the SDMT, where participants will be shown a total of 45 faces (15 each of old happy, old neutral or novel neutral faces) and asked to indicate whether the person shown is coming to the party or not (acceptance or rejection outcome), or whether they had not seen the person during the task.

Procedure

 Participants will first complete a demographic questionnaire followed by a battery of pre-experiment questionnaires on the laboratory computer including the sMFQ, SDQ and the AMT. After, participants will be given verbal instructions of the imagined scenario for the SDMT followed by written instructions on how to complete the task. Once the written instructions have been read, participants will complete two sample trials before commencing the actual SDMT. After completing the SDMT and related recall tasks, participants will be debriefed on the present study and recompensed with course credit for their participation.

Results

 A mixed model analysis will be used to test our first hypothesis that indivdiuals with high compared to low depression will show greater bias toward negative compared to positive memory cues. The outcome type (positive or negative) will be the within-subject factor and levels of depressive symptoms (high, moderate, low) will be the between-subjects factor with the rate of optimal choice made as the dependent variable.

To address our second hypothesis, a correlation analysis will be run to investigate whether levels of depressive symptoms are associated with corresponding extent of negative memory cue bias measured by the difference in percentage of optimal choice made for positive and negative outcome types on the SDMT for each participant. 

Prior to testing our last hypothesis, we will first use a general linear model to confirm that the groups with higher depressive symptoms show less memory specificity, where the levels of depressive symptoms will be the fixed factors and the percentage of specific memories on the AMT is the dependent variable. To test our third hypothesis that this difference is moderated by AMS, we will run a moderation analysis will be run in R using the lavaan package.

Consistent with existing literature, we firstly expect to see an overall diagnosis main effect (high vs. moderate vs. low symptom groups) where participants with higher scores on the sMFQ and SDQ will choose the optimal outcome significantly less on average than those with low symptoms (see Figure 2). We also expect those with higher depression to choose more optimal outcomes on average in trials with negative compared with positive memory cues. That is, we should observe a significant interaction effect between memory cue type and diagnosis main effects (see Figure 2). Further, we anticipate the extent of this bias for negative cues to have a significant positive correlation to levels of depressive symptoms. Thus, greater bias should predict a higher depressive symptom presentation score (see Figure 2).

Figure 2. Sample data displaying rate of optimal outcome across different level groups of depressive symptoms for memory cue types and expected significance in relation to one another.

Importantly, the results should show that in the high depressive symptom group, those with low AMS or low percentage specificity score on the AMT have lower rate of optimal outcome choice on the SDMT than those with high AMS.

Significance & Innovation

 Despite the expanding knowledge of depressive symptoms in the literature, there still remains uncertainties about the etiological foundations of, or what moderates their associated cognitive characteristics of memory biases and decision making skill deficits (Okwumabua, Duryea, & Wong, 2013). Ascertaining whether this association can be predicted by extent of negative memory bias or is directly moderated by AMS will be crucial for refining treatments and prevention methods that will aid indivdiuals’ emotional and mental well-being in the social domain. As aforementioned, there is already growing evidence for the effectiveness of positive and specific memory training intervention for depressed indivdiuals (Dalgleish & Werner-Seidler, 2014). The present study could therefore not only have practical but theoretical implications for future research to point towards a causative factor of depressive symptoms and traits. Further, testing whether the extent of bias toward negative memory and stimuli in social decision making can be used to predict levels of depressive symptoms will aid identify further avenues of enhancement in diagnostic and early-prevention methods (Gotlib & Joorman, 2010; Dalgleish & Werner-Seidler, 2014).

The present study addresses the association between memory biases and the social aspect of decision making in depression that has yet to be investigated in the literature. Previous research (Segrin, 2000; Wang et al., 2014; Hinterbuchinger et al., 2018) indicate predictive connections between depressive symptoms and social skill deficits in clinical and developmental samples, whilst not testing for the underlying influence of memory presentation and maintenance patterns distinct to depression on this link. Considering that depression is the most common mental health condition with 11% lifetime prevalence worldwide, the abnormal social decision making behaviour observed in those with MDD within social interaction contexts is an important area of research for examining predictors and risk factors of depression (Wang et al., 2014; Lim et al., 2018). Thus, the current study will add to the existing findings by exploring the link between the social decision making, memory bias and depression literature in the search for causal and diagnostic markers of depressive symptoms.

References

• Allen, N. B., & Badcock, P. B. T. (2003). The social risk hypothesis of depressed mood: Evolutionary, psychosocial and neurobiological perspectives. Psychological Bulletin, 129(6), 887-913.
• Askelund, A. D., Schweizer, S., Goodyer, I. M., & van Harmelen, A. (2019). Positive memory specificity is associated with reduced vulnerability to depression. Nature Human Behaviour, 3, 265-273.
• Bornstein, A. M., Khaw, M. W., Shohamy, D., & Daw, N. D. (2017). Reminders of past choices bias decisions for reward in humans. Nature, 8, 15958.
• Conway, M., A & Pleydell-Pearce, C. W. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107(2), 261-288.
• Dalgleish, T., & Werner-Seidler, A. (2014). Disruptions in autobiographical memory processing in depression and the emergence of memory therapeutics. Trends in Cognitive Sciences, 18(11), 596-604.
• Duncan, K. D., & Shohamy, D. (2016). Memory states influence value-based decisions. Journal of Experimental Psychology: General, 145(11), 1420-1426.
• Gotlib, I., & Joorman, J. (2010). Cognition and depression: Current status and future directions. Annual Review of Clinical Psychology, 6, 285-312.
• Gutenbrunner, C., Salmon, K., & Jose, P. E. (2018). Do overgeneral autobiographical memories predict increased psychopathological symptoms in community youth? A 3-year longitudinal investigation. Journal of Abnormal Child Psychology, 46, 197-208.
• Hinterbuchinger, B., Kaltenboarck, A., Baumgärtner, J. S., Mossaheb, N., & Friedrich, F. (2018). Do patients with different psychiatric disorders show altered social decision-making? A systematic review of ultimatum game experiments in clinical populations. Cognitive Neuropsychiatry, 23(3).
• Hitchcock, C., Gormley, S., O’Leary, C., Rodrigues, E., Wright, I., Griffiths, K., …Dalgleish, T. (2018). Study protocol for a randomised, controlled platform trial estimating the effect of autobiographical memory flexibility training (MemFlex) on relapse of recurrent major depressive disorder. BMJ Open, 8(1), 1-8.
• Kumar, P., Waiter, G. D., Magda, D., Milders, M., Reid, I., & Steel, J. D. (2017). Increased neural response to social rejection in major depression. Depression and Anxiety, 34(11), 1049-1056.
• Kuyken, W., & Dalgleish, T. (2011). Overgeneral autobiographical memory in adolescents at risk for depression. Memory, 19(3), 241-250.
• Lim, G. Y., Tam, W. W., Lu, Y., Ho, C. S., Zhang, M. W., & Ho, R. C. (2018). Prevalence of depression in the community from 30 countries between 1994 and 2014. Scientific Reports, 8(1), 2861. https://doi.org/10.1038/s41598-018-21243-x
• Matsumoto, N., & Mochizuki, S. (2018). Is retrieving positive self-concordant specific memory beneficial to mood repair? The moderation effect of depressive symptoms. Motivation and Emotion, 42, 276-286.
• McBride, C., Segal, Z., Kennedy, S., & Gemar, M. (2007). Changes in autobiographical memory specificity following cognitive behaviour therapy and pharmacotherapy for Major Depression. Psychopathology, 40(3), 147-152.
• Murty, V. P., FeldmanHall, O., Hunter, L. E., Phelps, E. A., & Davachi, L. (2016). Episodic memories predict adaptive value-based decision-making. Journal of Experimental Psychology: General, 145(5), 548-558.
• Neshat-Doost, H. T., Dalgleish, T., Yule, W., Kalantari, M., Ahmadi, S. J., Dyregroc, A., & Jobson, L. (2013). Enhancing autobiographical memory specificity through cognitive training: An intervention for depression translated from basic science. Clinical Psychological Science, 1(1), 84-92.
• Okwumabua, J. O., Duryea, E. J., & Wong, S. P. (2002). Decision making correlates of depressive symptoms among African-American adolescents: Implications for prevention approaches. American Journal of Health Education, 33(1), 28-35. https://doi.org/10.1080/19325037.2002.10604717
• Park, R. J., Goodyer, I. M., & Teasdale, J. D. (2002). Categorical overgeneral autobiographical memory in adolescents with major depressive disorder. Psychological Medicine, 32, 267-276.
• Ridout, N., Astell, A. J., Reid, I. C., Glen, T., & O’Carroll, R. E. (2003). Memory bias for emotional facial expressions in major depression. Cognition and Emotion, 17(1), 101-122.
• Rohde, L. A. (2016). Commentary: Transdiagnostic neuroscience of child and adolescent mental disorders – differentiating decision-making in attention-deficit/hyperactivity disorder, conduct disorder, depression and anxiety. A commentary on Sonuga-Barke et al. (2016). Journal of Child Psychology and Psychiatry, 57(3), 350-352.
• Segrin, C. (2000). Social skills deficits associated with depression. Clinical Psychology Review, 20(3), 379-403. https://doi.org/10.1016/S0272-7358(98)00104-4
• Wang, Y., Zhou, Y., Li, S., Wang, P., Wu, G. W., & Liu, Z. N. (2014). Impaired social decision making in patients with major depressive disorder. BMC Psychiatry, 14.