Neural Spike Shapes May Encode Recent Input History
A new preprint study suggests that the shape of a neuron's electrical pulse, or action potential, carries more information than previously thought. While standard neuroscience treats these signals as simple binary events (on or off), researchers from UC San Diego found that the waveform's specific shape changes based on the neuron's recent electrical input. In mouse visual cortex neurons, a machine-learning model could distinguish between different types of stimulation with 78.4% accuracy based solely on the signal's shape. In rat brain recordings, 30 out of 40 neurons exhibited multiple distinct signal shapes. While this implies that current recording methods may discard valuable data about a neuron's recent history, the study does not yet prove that these shape variations are actively used to communicate with other neurons. The findings highlight a potential blind spot in how scientists interpret brain activity, though the data is limited to small sample sizes and requires peer review.
