Fig. 2.

Decoding monkeys’ identity based on spatial occupancy reveals structured pattern of spatial utilization. (A) Schematic of the decoding pipeline for monkeys’ identity from their spatial occupancy pattern. We trained the Support Vector Machine classifier (SVM) on the simultaneously recorded monkey positions using 3 dimensions. Using simultaneous positions across time recorded for each monkey pair we decoded monkeys’ identity such that each classifier was trained to distinguish between a pair of monkeys. (B) Example of one such trained classifier. The trained classifier was tested in a cross-validated manner (C) or on all remaining 18 recording sessions (D). (C) To decode spatial occupancy within each session and monkey pair, a cross-validation procedure was performed. Decoding accuracy was above the chance level for all monkey pairs (bottom left panel), showing that each animal sustained its spatial footprint within each session. (D) The same decoding was performed on data across sessions which showed a similar pattern of results. Left panel: Decoding accuracy was above the chance level for all monkey pairs, demonstrating that each animal sustained its spatial footprint across multiple days. Right panel: Average accuracy scores for each monkey pair are summarized in matrix form in the bottom right panel.