FReD is introduced, which derives fMRI representations from a frozen Deep Compression AutoEncoder pre-trained exclusively on natural images and pairs them with a task specific readout, making frozen natural-image features as a useful baseline for assessing its added value on current fMRI benchmarks.
Abstract
Foundation models pre-trained on large-scale fMRI datasets have shown strong downstream performance, but at substantial data and computation cost. To investigate how much fMRI-specific pre-training is actually needed for such performance, we introduce FReD, which derives fMRI representations from a frozen Deep Compression AutoEncoder (DCAE) pre-trained exclusively on natural images and pairs them with a task specific readout. For trait prediction, FReD summarizes frame-wise representations by their temporal mean and log-standard deviation and applies linear probing, with late fusion across two normalization schemes. For state prediction, it represents each frame as a single token and models temporal dependencies with a shallow Transformer. Across four resting-state datasets spanning six trait-prediction targets, linear probes on frozen DCAE features generally outperform those on fMRI foundation model representations and remain competitive with fully fine-tuned fMRI foundation models. On three task-fMRI state-prediction tasks, a temporal readout on DCAE features performs comparably to the strongest foundation models evaluated. A Gaussian injection analysis further shows that localized signal changes are recovered more accurately from the frozen DCAE features than from the evaluated foundation-model representations. Together, these results show that strong performance on current fMRI benchmarks is possible without fMRI-specific representation pre-training, making frozen natural-image features as a useful baseline for assessing its added value.
It is shown that MnemoDyn is compute efficient and generalizes very well across diverse populations and scanning protocols, and when benchmarked against current state-of-the-art transformer-based approaches, MnemoDyn consistently delivers superior reconstruction quality.
Sourav Pal, V. Lương, Hoseok Lee et al.· 1 citation
These findings support the contribution of Transformer processing to language-map prediction and distinguish it from subject-specific spatial conditioning and relative uncertainty estimation.
D. Di Giovanni, D. L. Collins· bioRxiv· 0 citations
It is shown that KRR is weighted by the eigenvalues of the FC which are miscalibrated for phenotype prediction, and an efficient spectral filter is applied to recalibrate the eigenvalues of each subject's FC matrix, enabling the model to exploit more inter-individual variance.
G. Marraffini, Victoria Shevchenko, Carlo Alberto Barbano et al.· arXiv.org· 0 citations
Neural State Prediction (NSP), a latent-predictive framework that constrains both the prediction target and the available context, is introduced and jointly designing latent targets and their context offers a promising direction for EEG foundation models that learn from distributed signal structure.
Kieren Yu, Zi-Yang Liu, Chang Huang et al.· 0 citations
Findings support organizing fMRI pretraining and adaptation by measured learning relations rather than treating domains and tasks as independent flat sets.
Jun-Feng Xia, Wen-Hao Ye, Jun-Xiang Zhang et al.· 2 citations
A general fMRI sequence prediction model, the Frequency-Filtered Attention Transformer (FFAformer), which models low-frequency variations in the frequency domain to capture long-range dependencies and incorporates FC consistency constraints to preserve brain network structure.
Chengcheng Du, Fei-Fei Zhao, Yin-Qian Sun et al.· iScience· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.