Large-scale synaptic connectivity mapping with silicon chip recording

• Researchers mapped and catalogued more than 70,000 synaptic connections from about 2,000 rat neurons using a silicon chip capable of recording small synaptic signals from a large number of neurons. ¹
• The work advances neuronal recording methods and moves toward detailed synaptic connection maps of the brain. ¹
• The approach uses direct electrophysiology and neural recording on silicon. ^{1 1}

Weekly enrichment (2026-07-20)

• The peer-reviewed study was published in Nature Biomedical Engineering (2025; DOI 10.1038/s41551-025-01352-5) by Jun Wang, Woo-Bin Jung, Rona S. Gertner, Hongkun Park and Donhee Ham of the Harvard John A. Paulson School of Engineering and Applied Sciences.^{2 3}
• The device is a 4,096-electrode microhole array (a 64 × 64 pixel grid) fabricated on a complementary metal-oxide-semiconductor (CMOS) chip, with microholes etched into the chip and coated with platinum black to roughen the surface and improve the neuron interface.^{3 4}
• Intracellular access was obtained by electroporation (gentle current injection through the electrodes), and the array achieved a 90% average intracellular coupling rate (3,685 of 4,096 pixels), reaching 97% (3,988 of 4,096) in the best case.^{3 4}
• The microhole design produced roughly five times larger intracellular recording amplitude than the team’s earlier vertical nanoneedle electrode array, improving data quality.⁴
• From the network-wide recordings the team extracted more than 70,000 plausible synaptic connections among more than 2,000 cultured rat neurons, with an estimated overall error rate of about 5%.^{2 3}
• Connections were catalogued into four classes: electrical synapses plus inhibitory, weak/uneventful excitatory and strong/eventful excitatory chemical synapses, capturing both connectivity and connection strength.^{3 4}
• The result is a large jump over prior parallelized intracellular recording, which had a state-of-the-art mapping limit of about 300 connections (the group’s own 2020 nanoneedle device).^{2 4}
• The approach aims to combine the sensitivity of patch-clamp recording (which reveals connection strength) with the scale of extracellular multi-electrode arrays; the team reports it is now working toward a design deployable in a live brain, and the work was supported by the Samsung Advanced Institute of Technology.^{2 4}

Footnotes

1. https://www.sciencedaily.com/releases/2025/02/250213143254.htm ↩ ↩² ↩³ ↩⁴

2. https://doi.org/10.1038/s41551-025-01352-5 ↩ ↩² ↩³ ↩⁴

3. https://pubmed.ncbi.nlm.nih.gov/39934437/ ↩ ↩² ↩³ ↩⁴ ↩⁵

4. https://seas.harvard.edu/news/2025/02/mapping-connections-neuronal-network ↩ ↩² ↩³ ↩⁴ ↩⁵ ↩⁶