Neuroscience · Cognitive Neuroscience
Researchers found that information flow between major brain networks differed between autistic participants and non-autistic controls. An artificial-intelligence system classified the two groups with a mean accuracy of 91% across cross-validation, but the study does not establish a clinical biomarker.
Brain stimulation separated motion-discrimination accuracy from confidence in perceptual decisions. Strengthening one visual pathway affected both measures, while disrupting a parietal brain region changed confidence-related performance without altering motion discrimination.
Neural networks running on low-power microcontrollers decoded surface muscle signals in real time, a requirement for wearable devices and prosthetic control. On a large-scale benchmark, the models matched or exceeded earlier accuracy with up to 5,000 times lower computational cost in resource-constrained settings.
