Identifying aligned tensor decompositions across heterogeneous neural datasets

Neuralign identifies a single label-free component across 93 brain-wide mouse recordings that isolates activity linked to inhibition of the anterior lateral motor cortex (ALM).

• Why it matters: Neural recordings differ in which neurons and trials are sampled, making it hard to find shared structure without relying on matched samples or behavioural labels.
• What they did: Neuralign jointly aligns and decomposes heterogeneous recordings into components pairing shared temporal patterns with trial- and neuron-specific expression, using tensor-based modeling without matched neurons, trials, or labels.
• The result: A reference learned from correct, uninhibited trials shows inhibition-related activity outside the choice direction on error trials, strongest in ALM and thalamus.

AI summary of the abstract · may contain errors · not medical advice · read the paper.