Alexander Gray and Dr. Michael Larson, Psychology Department

Main Text

This study will help us to better understand the neurophysiological correlates of error processing–the cognitive function of detecting mistakes. Also, this study will help clarify the potential generation process of this neurophysiological correlate, specifically called the error-related negativity (ERN).

The ERN is a scalp-recorded, negative electrical potential detected approximately 50 ms after making a mistake compared to correct response and is generated from the anterior cingulate cortex (Dehaene, Posner, & Tucker, 1994). The ERN amplitude is consistently larger in individuals who have affective trait disorders such as obsessive-compulsive disorder (Gehring, Himle, & Nisenson, 2000), high negative affect (Luu, Collins, & Tucker, 2000), or mild depression (Schrijvers et al., 2008). This consistent manifestation has lead many to wonder if affect may play a role in ERN generation. As such, some research has examined the role of affective states on ERN generation and found mixed results. For example, individuals who received negative, derogatory feedback after making an error, and thereby inducing them into an anxious mood, demonstrate larger ERN amplitudes (Wiswede, Munte, & Russeler, 2009). However, individuals with arachnophobia and induced into a fear condition due to a tarantula in close proximity, demonstrated ERN amplitudes similar to controls(Moser et al., 2005). This study tested the hypothesis that ERN amplitudes would be modulated in similar directions for induced mood states as they are for mood-like trait conditions. Specifically, ERN amplitudes would increase in anxiety and sad mood states and decrease in happy and calm mood states.

This study had 138 participants and 37 were excluded due to noise interfering with ERN detection, insufficient error commission, or disqualifying depression scores or medication histories. Of the 101 remaining participants, 25 were in the calm condition, 27 in the anxious condition, 26 in the happy condition, and 23 in the sad condition. Participants were first screened for depression, recreational drug use, concussions, and asked if they had any previous psychological or neurological illnesses. After this exclusion, all participants compared did not significantly differ by state or trait anxiety or positive or negative affect.

Then after connecting participants to an electroencephalogram (EEG), which passively detects post-synaptic potentials in the brain, participants were instructed on the Flanker task. This task consists of identifying the correct direction of the middle of five arrows in congruent (i.e., <<<<< or >>>>>) and incongruent (<<><< or >><>>) trials for 600 trials. This task elicited some erroneousness responses that would elicit the ERN.

Before beginning the Flanker, participants rated their baseline mood on an affect grid in which they determined the valence (pleasantness from very high to very low) and the arousal (energy level from very high to very low) of their mood. Then they were induced into one of the four mood conditions by listening to mood appropriate classical music and ruminating on mood appropriate past events for ten minutes. For example, an individual who was asked to be anxious would listen to jarring musical selections from Mars the Bringer of War while reflecting on personal, past incidences such as failing a test or breaking up a significant relationship. Participants rated their mood five minutes into the mood induction procedure, after the mood induction procedure, and three times during the Flanker task.

With the collected data, we compared successive mood states to baseline using a paired sample t-test. We found that for valence and arousal, calm states did not significantly differ from baseline and happy participants’ mood only differed from baseline during the mood induction procedure. However, participants mood did differ from baseline during most rating points for valence and arousal in sad and anxious mood conditions.

Then we conducted a repeated measures analysis of variance (ANOVA) on four variables: ERN, latency (when the ERN occurs), response time (when the participants responds to a stimulus), and error rate. A 4-Group x 2-Congruency ANOVA revealed that mood groups did not differ in error rates or response time but did differ by trial congruency. This is consistent with literature. Also a 4-Group x 2-Accuracy ANOVA revealed that groups did not differ in latency or ERN amplitude, contrary to our prediction.

One possible explanation to account for finding the null is that our mood induction procedure did not reliably work for calm and happy mood groups and only mostly worked for anxious and sad groups. Mood generally returned to baseline as the Flanker task continued and previous research (Luu, Collins, & Tucker, 2000) suggests that ERN amplitudes are attenuated with the length of the task. This may be a sign of participant fatigue or growing disinterest. In the near future, we expect to compare ERN signals from the first 200 trials of all participants to see if ERN amplitudes will significantly differ across groups.
In addition, our findings may be due to the high exclusion rates due to low participant error rates, EEG noise, or participants met exclusionary criteria. Also, we choose to exclude participants who made 5 errors or less when standard protocol suggests that participants who make 6 errors or less are excluded (Olvet & Hajcak, 2009). While this can influence the data, we needed to keep our power high enough to detect a difference if there was a difference.

At last, our data may suggest an alternative framework for ERN generation. ERN amplitudes may not be differentially modulated by affect as they may be by the importance of error commission (Olvet & Hajcak, 2008). Several studies support this theory. For example, in OCD studies, participants were naturally more hyper-vigilant and concerned in making correct answers than controls were (Gehring et al., 2000). Individuals with mild depression may be more sensitive to social cues for failure or fear the possibility of being negatively evaluated (Schrijvers et al., 2008). Individuals who receive negative feedback after error commission will attempt to improve to avoid future feedback (Wiswede et al., 2009). These studies included paradigms in which error commission may have been significant to the participants. However, in the spider-phobic study the spider was present regardless of response choice and thus erroneous choices could not have been significant. Similarly, in our study participants were in a mood state regardless of response choice and erroneous choices would not have been more significant compared to other groups.

Reference

• Dehaene, S., Posner, M. I., & Tucker, D. M. (1994). Localization of a neural system for error detection and compensation. Psychological Science, 5, 303-305.
• Luu, P., Tucker, D. M., & Collins, P. (2000). Mood, personality, and self-monitoring: Negative affect and emotionality in relation to frontal lobe mechanism of error monitoring. Journal of Experimental Psychology: General, 129, 43-60.
• Moser, J. S., Hajcak, G., & Simons, R. F. (2005). The effects of fear on performance monitoring and attentional allocation. Psychophysiology, 42, 261-268.
• Olvet, D. M. & Hajcak, G. (2008). The error-related negativity (ERN) and psychopathology: Toward an endophenotype. Clinical Psychology Reviews, 28, 1343-1354.
• Olvet, D. M., & Hajcak, G. (2009). The stability of error-related brain activity with increasing trials. Psychophysiology, 46, 957-961.
• Schrijvers, D., de Bruijn, E. R. A., Mass, Y., De Grave, C., Sabbe, B. G. C., & Hulstijn, W. (2008). Action monitoring in major depressive disorder with psychomotor retardation. Cortex, 44, 569-579.
• Wiswede, D., Munte, T. F., & Russeler, J. (2009). Negative affect induced by derogatory verbal feedback modulates the neural signature of error detection. Social Cognitive and Affective Neuroscience, 4, 227-237.