Model-based and model-free valuation signals in the human brain vary markedly in their relationship to individual differences in human behavioral control
Bibliographic record
Abstract
Human action selection under reinforcement is thought to rely on two distinct strategies: model-free and model-based reinforcement learning. While behavior in sequential decision-making tasks often reflects a mixture of both, the neural basis of individual differences in their expression remains unclear. To investigate this, we conducted a large-scale fMRI study with 179 participants performing a variant of the two-step task. Using both cluster-defined subgroups and computational parameter estimates, we found that the ventromedial prefrontal cortex encodes model-based and model-free value signals differently depending on individual strategy use. Model-based value signals were strongly linked to the degree of model-based behavioral reliance, whereas model-free signals appeared regardless of model-free behavioral influence. Leveraging the large sample, we also addressed a longstanding debate about whether model-based knowledge is incorporated into reward prediction errors or if such signals are purely model-free. Surprisingly, ventral striatum prediction error activity was better explained by model-based computations, while a middle caudate error signal was more aligned with model-free learning. Moreover, individuals lacking both model-based behavior and model-based neural signals exhibited impaired state prediction errors, suggesting a difficulty in building or updating their internal model of the environment. These findings indicate that model-free signals are ubiquitous across individuals, even in those not behaviorally relying on model-free strategies, while model-based representations appear only in those individuals utilizing such a strategy at the behavioral level, the absence of which may depend in part on underlying difficulties in forming accurate model-based predictions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.004 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".