- Brain‑wide decision maps
A flagship collaboration (International Brain Laboratory) recorded from > 600,000 neurons across ~279 brain regions in mice during decision‑making tasks. Their findings challenge modular views: decision signals are broadly distributed, with sensory, motor, and associative areas all participating. - Structure–function coupling & parcellation issues
A review in Nature Reviews Neuroscience examines methodological pitfalls in how brain parcellation choices influence estimates of structure–function coupling (i.e. how anatomical connectivity constrains functional dynamics). The authors argue for more principled parcellation strategies to avoid biased coupling metrics. - Nanoscale connectomics & network neuroscience
A recent conceptual review urges that network neuroscience should lean more heavily into nanoscale connectomic data (synapse‑level, cellular annotations) rather than relying solely on meso‑ or macroscale abstractions. This more granular scale enables mechanistic interpretability. - Causal frameworks for computational neuroscience
An up‑to‑date review argues that adopting formal causal inference perspectives (e.g. directed acyclic graphs, intervention logic) can sharpen experimental design and data analysis in neuroimaging and electrophysiology, mitigating confounds like selection bias or latent variables. - Memory‑augmented Transformers bridging neuroscience and ML
A systematic review links principles from biological memory (e.g., multi‑timescale buffers, consolidation, gating) to architecture designs in memory‑augmented Transformer models, charting paths toward better context retention, lifelong learning, and knowledge integration.
🧠 Neuroscience & Neural Theory
Discover more from ExNTER.com
Subscribe to get the latest posts sent to your email.
Leave a Reply