Neural correlates of the subjective experience of free-will during value-based risky decisions: a pilot study
neuroscience
Quick verdict
fMRI pilot (n=12): free-will ratings in risky decisions track mid-cingulate/MFG and hippocampal BOLD; intriguing hypothesis-generating work but underpowered.
Summary
This bioRxiv preprint reports a pilot fMRI study asking whether BOLD activity tracks the subjective experience of free-will during value-based decisions. Twelve participants chose between gambles and sure things in a task condition (choose the option that pays more) and a matched control condition (choose the instructed option, which replicated each participant's own task-block choices). After each block, participants rated experienced free-will on a 1-5 scale via an anti-anchoring interface with randomized motor mappings. Ratings were higher for task than control (mean +1.12, p=0.041). Whole-brain correlations of the (Task-Control) contrast with rating differences showed positive associations in mid-cingulate/BA32 (FWE p=0.033) and left middle frontal gyrus (p=0.006), and a negative association in left hippocampus/parahippocampal gyrus (p=0.052); clusters defined from the n=30 main-effect analysis, including precentral/IFG and insula, also correlated with ratings. Compared with 18 participants who never rated free-will, the reporting group showed trend-level striatal/putamen hyperactivation for (Task-Control), interpreted as introspective monitoring. Ratings showed no relation to risk, conflict, consistency, or difficulty; only a positive RT trend emerged (r=0.46, p=0.134). Key limitations: small sample, only three retrospective block-level ratings per participant, control block always last, SVM-predicted scanner start times for four participants, and ROI analyses whose defining sample (n=30) overlaps the tested sample. The authors appropriately frame the results as hypothesis-generating.
Review complete
Assessment at a glance
5
risk flags identified
● high 0
● medium 4
● low 1
5 strengths · 5 weaknesses · 5 author questions
Levels reflect verification complexity — not a quality score. Evolve never rates your paper.
Key claims under review
- Self-reported experienced free-will is higher when choosing to maximize payoff than when following a matched instruction in the same risky-choice trials (mean difference 1.12, p=0.041).
- Decision-related BOLD (Task-Control) in mid-cingulate/BA32 and left middle frontal gyrus correlates positively across participants with free-will rating differences, while left hippocampus/parahippocampal gyrus correlates negatively.
- Neural correlates of subjectively experienced free-will do not coincide with the main-effect contrasts of free vs. instructed choice, echoing Filevich et al. (2013).
- The mere requirement to later report free-will is associated with trend-level elevated striatal/putamen (Task-Control) activation relative to non-reporting participants, suggesting pre-emptive introspection or monitoring.
- Behaviorally, free-will ratings relate only weakly to performance: a positive RT trend (r=0.46, p=0.134) with no association to risk, conflict, strategy consistency, or difficulty.
Strengths
Clear conceptual reframing
The Introduction and Discussion carefully separate the subjective experience of free-will from the ontological debate and from objectively-defined free-choice contrasts (instructed vs. own choice), building on Filevich et al. 2013 and Rens et al. 2018 while extending the question to value-based economic decisions, where they claim a first.
Thoughtful rating-interface design
The rating screen starts from an invalid value drawn from {0, 2.5, 3.5, 6} that must be actively adjusted, with randomized initial values and randomized finger-to-action mappings per block, explicitly decoupling the rating from anchoring and differential motor effort (Methods, Free-will ratings; Figure 2).
Stimulus-matched control condition
Transparency about weak effects and selection bias
Manipulation check on the reporting requirement
Weaknesses
Very noisy key variable at n=12
Partial circularity in Table 2 ROI selection
Fixed block order confound
SVM-imputed scanner start times
Reporting and availability gaps
Full Review Report
$1.99 · one-timeUnlock the complete preliminary analysis Evolve produced for this paper.
- All 5 strengths & 5 weaknesses, analyzed in detail
- 5 risk flags with severity, section & reviewer questions
- Methods & statistical evidence assessment
- 5 questions a peer reviewer would ask
- Full six-dimension evaluation report & recommendations
One-time purchase · this paper only
One-time purchase, this paper only. AI-generated preliminary analysis — not peer review.
Frequently asked questions
What brain regions are linked to the feeling of free will in this study?
In this pilot fMRI study of 12 participants, stronger self-reported free-will during risky decisions was associated with larger task-versus-control BOLD differences in mid-cingulate cortex (BA 24/32) and left middle frontal gyrus, and with smaller differences in left hippocampus and parahippocampal gyrus. Bilateral insula correlations appeared after small-volume correction within independently defined clusters.
Does this study prove that free will exists?
No. The authors explicitly avoid the philosophical debate about whether free will exists. They study the subjective experience of free-will, i.e., how strongly people feel they chose freely, and map that feeling onto brain activity and behavior during a value-based decision task.
Did being asked to report free-will change participants' brain activity or choices?
Possibly. Comparing 12 participants who rated free-will with 18 who did not, the rating group showed trend-level higher striatal (putamen/pallidum) task-control activation, which the authors interpret as monitoring or pre-emptive introspection. Because no rating group data exist for the non-raters, the effect of reporting on the experience itself could not be assessed.
How reliable are the results given the sample size?
The study had only 12 participants, each giving three retrospective block-level ratings. The authors themselves call it a pilot with alpha-inflation risk and state that replication with a larger cohort is needed; several key effects were trend-level only (p between 0.05 and 0.14), so the findings should be treated as hypothesis-generating.
About this review
This is an AI-generated preliminary review by Evolve.Pub, produced from the openly available text of the paper. It is not peer review and does not represent the authors' views. The canonical record of the paper is at the publisher link above.
- Figures 1-4 are available only as captions; plotted data such as rating distributions and contrast bar plots could not be verified from the text.
- Supplementary material (gamble set, payment details, tables S1-S4) and the companion preprint (Modak and Brown 2026) were not embedded; several methods details depend on them.
- Minor extraction artifacts: repeated license boilerplate interleaves the text, and the striatum result is referenced as 'figure 2' but captioned 'Figure 4', an apparent typo in the manuscript.
- No data or code availability statement was found in the extracted text; codeAvailable=false reflects absence of a statement rather than a confirmed refusal to share.
Go deeper on this paper
Wrote a paper? Stress-test it before reviewers do.
Upload your manuscript for the same structured scrutiny — reasoning, statistics, evidence, figures, and reproducibility.