Neural Latent Aligner: Cross-trial Alignment for Learning Representations of Complex, Naturalistic Neural Data
This work addresses the challenge of analyzing complex, low-SNR neural data in neuroscience, offering a method to improve representation learning for decoding behaviors, though it appears incremental as it builds on existing alignment and time warping techniques.
The authors tackled the problem of learning behaviorally relevant neural representations from complex, noisy neural data by proposing Neural Latent Aligner (NLA), an unsupervised framework that aligns representations across repeated trials and includes a differentiable time warping model (TWM) to handle temporal misalignment. When applied to intracranial ECoG data from natural speaking, the model learned better representations for decoding behaviors than baselines, especially in lower-dimensional spaces, and demonstrated improved cross-trial consistency and shared neural trajectories.
Understanding the neural implementation of complex human behaviors is one of the major goals in neuroscience. To this end, it is crucial to find a true representation of the neural data, which is challenging due to the high complexity of behaviors and the low signal-to-ratio (SNR) of the signals. Here, we propose a novel unsupervised learning framework, Neural Latent Aligner (NLA), to find well-constrained, behaviorally relevant neural representations of complex behaviors. The key idea is to align representations across repeated trials to learn cross-trial consistent information. Furthermore, we propose a novel, fully differentiable time warping model (TWM) to resolve the temporal misalignment of trials. When applied to intracranial electrocorticography (ECoG) of natural speaking, our model learns better representations for decoding behaviors than the baseline models, especially in lower dimensional space. The TWM is empirically validated by measuring behavioral coherence between aligned trials. The proposed framework learns more cross-trial consistent representations than the baselines, and when visualized, the manifold reveals shared neural trajectories across trials.