Active-Voxel Selection for Small-Sample fMRI Decoding: Dataset-Dependent Preprocessing and the Cost of Cross-Validation Leakage
Abstract
Abstract Decoding what a person is seeing or thinking from functional MRI typically has to be done from a few dozen to a few hundred labeled brain volumes, each containing tens of thousands of voxels. In this small-sample regime, how voxels are selected and how cross-validation is organized can matter as much as the choice of classifier. We generalize an active-voxel selection method developed for binary picture-versus-sentence decoding (Chowdhury, 2014) to multi-class decoding and evaluate it under fully nested cross-validation on two public datasets: picture-versus-sentence decoding in the StarPlus data of Mitchell et al. (three subjects) and eight-way object-category decoding from ventral temporal cortex in the Haxby et al. data (six subjects). The method reaches 92.9% mean accuracy on StarPlus, against 85.0% for the active-ROI method of Chowdhury (2014). On Haxby our leakage-free pipeline (ventral temporal mask, feature standardization, tuned regularization, nested leave-one-run-out evaluation) reaches 71.8% using every ventral temporal voxel and 71.4% with the active-voxel selection step added, far above the run-grouped version of nilearn's public Haxby example (37.2%), a gap that restricting the example's own pipeline to the ventral temporal mask closes almost entirely by itself. Within the anatomical region, therefore, the selection step itself adds nothing (the no-selection variant coincides with the field's standard whole-VT analysis), and the inner loop most often…
Read the paper
Related studies
- Comparative effectiveness of GLP-1 receptor agonists on glycaemic control, body weight, and lipid profile for type 2 diabetes: systematic review and network meta-analysis
- Adverse effects of cyclooxygenase 2 inhibitors on renal and arrhythmia events: meta-analysis of randomized trials
- Benefits and harms of drug treatment for type 2 diabetes: systematic review and network meta-analysis of randomised controlled trials
- Comparative effectiveness and safety of statins as a class and of specific statins for primary prevention of cardiovascular disease: A systematic review, meta‐analysis, and network meta‐analysis of randomized trials with 94,283 participants
- Screen time increases overweight and obesity risk among adolescents: a systematic review and dose-response meta-analysis
- Exposure to WHO AWaRe antibiotics and isolation of multi-drug resistant bacteria: a systematic review and meta-analysis
Go deeper in Evidence Hub
Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.