Work / CABLE / CoSyNE 2027 submission

A linear model that outperformed deep-learning baselines on new people.

What if the EEG-to-fMRI bottleneck was aligning individuals, rather than finding a larger neural network? CABLE tested that question with subject-specific normalization and a shared linear model.

0.41Correlation, CABLE
0.30NeuroBOLT baseline
+37%Relative improvement
22/22Held-out subjects ahead
01 / How the method works

Align the subject, then make the prediction.

Each EEG recording is whitened using its own running, past-only covariance. A regularized linear readout then maps those aligned signals to regional BOLD responses. The method uses no labeled fMRI calibration for a new person.

Full uncropped CABLE pipeline diagram from scalp EEG to subject-specific alignment to regional BOLD response
Pipeline diagram from the CoSyNE submission.

Alignment

Reduce subject-specific sensor differences using earlier EEG observations.

Linear model

Read out regional brain signals with lags and ridge regression.

Practical efficiency

Train in CPU minutes on the tested datasets rather than requiring GPU-hour runs.

02 / Prespecified evaluation

New people, and a new recording site.

The main comparison uses 22 people with identical subject-disjoint evaluation splits across CABLE, MEL, and NeuroBOLT. The manuscript reports r = 0.41 for CABLE, r = 0.35 for MEL, and r = 0.30 for NeuroBOLT. An independent dataset of 33 adults provides an additional external-site test.

CABLE comparison across fresh splits, leave-one-out tests, external subjects and regional BOLD signals
Benchmark panels, including the external-dataset comparison.
03 / What produced the gain

The alignment step did most of the work.

The ablation study contrasts raw spectral features, different alignment stages, the final ensemble, and larger feature families. More nonlinear capacity added little in these tests. It is a reminder that sometimes the useful insight is a better representation of the data.

Separate figure showing ablation, data efficiency, and person-specific error
Ablation, data efficiency and remaining subject-specific error. These are results from the submitted paper, not clinical validation.
Research and publication record ↗