Response Monitoring on a Face Processing Task and Its Relation to the ERN

Background:

Individuals with High-Functioning Autism (HFA) have difficulty with executive functioning skills, including response monitoring.

Objectives:

1) To compare HFA and control participants on behavioral indices of response monitoring during a face processing task.

2) To examine the relation between response monitoring on a face processing task and a neurophysiological index of response monitoring, the Error-Related Negativity (ERN).

Methods:

Twenty children with HFA and 16 children with typical development participated in this study. Participants were presented with a series of upright and inverted faces that were revealed in piecemeal. Participants were instructed to guess the affect of the face as quickly and accurately as possible. After the whole face was revealed, participants could change their original affect selection. The two dependent variables of interest were: 1) percentage of trials in which the original affect selection was correct and 2) percentage of trials in which the participant correctly kept the original affect or correctly selected a new affect. EEG data were collected using Lycra stretch Electrocaps as participants completed a modified Flanker task. Participants were instructed to push a button to identify the direction of the middle arrow on compatible (<<<<< or >>>>>) and incompatible (<<< or >>>) trials. ERN was quantified as the maximum negative peak within 150 ms of the error response. Data were analyzed from site Fz.

Results:

There was an interaction between diagnostic group and orientation on original affect selection, F(1, 32) = 4.61, p = 0.04, η²_p = 0.13. Post hoc analyses revealed that participants with autism were marginally more accurate on selecting affect for inverted faces than control participants, F(1, 32) = 3.73, p = 0.06, η²_p = 0.10. There was also an interaction between diagnostic group and face orientation on the decision to keep/change the affect, F(1, 32) = 4.05, p = 0.05, η²_p = 0.11. Post hoc analyses revealed that control participants made better decisions for upright faces than participants with autism, F(1, 32) = 5.85, p = 0.02, η²_p = 0.16.

A regression was performed with age, diagnostic group, and ERN amplitude as predictors of a composite score of the two face processing variables. Age was a significant predictor of face processing, B = 0.46, t(32) = 3.15, p < 0.01, such that older individuals performed better on the task. Over and above the effects of age, ERN amplitude significantly predicted face processing performance, B = -0.31, t(32) = -2.16, p = 0.04, such that individuals with a higher ERN amplitude had better face processing.

Conclusions:

Consistent with the literature, individuals with autism didn’t show the typical advantage for processing upright faces and disadvantage for processing inverted faces. Of particular interest, individuals, regardless of diagnostic group, with heightened error-monitoring on a non-social, Flanker task were better at error-monitoring on a social, face processing task. Response monitoring may be a helpful venue for intervention in autism, as it appears to be an integral component of complex social information processing.