Strategies to Decipher Neuron Identity from Extracellular Recordings in Behaving Nonhuman Primates.

Area of Science:

• Neuroscience
• Computational Neuroscience
• Primate Research

Background:

• Accurate neuron type identification is crucial for understanding neural circuit computation using extracellular recordings in behaving animals.
• Current recording probes have limitations in resolving neuron identity, hindering detailed circuit analysis.

Purpose of the Study:

• To develop a generalizable framework for assigning neuron types from extracellular recordings in nonhuman primates.
• To leverage deep learning models trained on expert-identified neurons for automated neuron classification.

Main Methods:

• A framework combining logic, circuit architecture, laminar information, and functional discharge properties was developed.
• Expert identification of neuron types in rhesus macaques during smooth pursuit eye movements was performed.
• Deep-learning classifiers were trained using extracellular features like waveform, discharge statistics, and anatomical layer information.

Main Results:

• Waveform, discharge statistics, and anatomical layer provided significant information for neuron identification.
• Deep-learning classifiers integrating these features achieved enhanced neuron identification accuracy.
• The methodology was validated in the cerebellar floccular complex during smooth pursuit eye movements.

Conclusions:

• The developed framework and deep-learning tools enable accurate neuron type identification from extracellular recordings.
• This generalized methodology supports characterization of information processing within neural circuits across species.
• It lays essential groundwork for advancing our understanding of neural computation.