Neuronal XOR computation requires temporally activated synaptic information
Single neurons can perform complex exclusive-or (XOR) computation which is conventionally thought to need multilayer networks, yet the biophysical mechanisms and the robustness of such computations under stochastic synaptic drive remain unclear. Using biologically realistic modeling, we demonstrate that dendritic depolarization block provides a universal and robust mechanism for XOR-like computation. Voltage-gated calcium channels enable a non-monotonic input-output relationship: moderate asynchronous drive evokes spiking, whereas excessive drive induces a sustained plateau that suppresses output, naturally implementing XOR logic computation. This computation is strictly gated by temporal distribution of synaptic inputs rather than spatial distribution; the XOR-like response is robust across diverse dendritic configurations but collapses under synchronous activation. Dendritic depolarization block propagates to the soma via dendrite-soma coupling, establishing a neuronal output-limiting mechanism. The computation is highly efficient, leveraging a frequency-synapse-count trade-off to ensure stable logic operations across varying levels of background activity. Cross-morphological validation confirms that XOR-like computation is morphology-independent, representing a generalizable biophysical principle of neural information processing.
Authors
- Yunliang Zang (ORCID: https://orcid.org/0000-0001-8999-1936)
- Xiaojuan Sun
- Yanheng Li
Institutions
- Beijing University of Posts and Telecommunications (CN)
- Tianjin University (CN)
Publication Details
- Journal
- PLoS Computational Biology
- Published
- 2026-11-16
- DOI
- https://doi.org/10.1371/journal.pcbi.1014971
- Primary Topic
- Neural dynamics and brain function
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China
- National Key Research and Development Program of China