Here we use and extend the classic distinction between instructed and free choice to investigate the neural correlates of subjective voluntariness.
This tradition has therefore capitalised on objective experimental design factors which allow instructed and free choice to be defined in terms of information that is provided by external cues, or is not so provided.
In this free-viewing experiment, participants were instructed to free-view images of natural scenes presented on a 21-inch CRT monitor at a viewing distance of 0.75 m while their eye movements were recorded by an eye tracking apparatus.
Experimental tasks have typically studied free choice by contrasting free and instructed selection of response alternatives.
Results of the comparison between free and instructed conditions showed significantly greater BOLD responses for free compared with instructed in RCZ/SMA, left DLPFC, bilateral IPL and left PMC.
If the assumption is correct, and there is a correspondence between objective and subjective accounts of free action, then the neural correlates of free and instructed choices defined objectively should roughly match with those defined subjectively.
Second, in an objective context, participants performed the free and instructed conditions again, but each trial was followed by a rating of how free they felt their immediately-preceding action choice had been.
Specifically, in those areas identified by classical free > instructed choice, we found no evidence for stronger activity when participants felt subjectively more free compared to when they felt subjectively less free.
Our main aim was to compare the brain correlates of the extreme free and instructed conditions with those of a subjective context in which participants could themselves report how free their choice had been by means of a VAS (see below).
Therefore, to further examine the mechanisms associated with the feeling of voluntariness, and their relationship to the objective operationalization of free action, we did a whole-brain analysis for the contrast free > instructed (median split) in the subjective context controlling for stimulus space and RT by including them as regressors in the first-level model.
We measured BOLD activations in classical conditions in which these free and instructed actions were defined in the objective, classical way, either leaving the response choice to the participant (free trials), or fully specifying the required response in each trial (instructed trials).
