Fig. 3: Encoding of behaviors induced by the threatening object.

a Schematic of Object-threat test. During habituation, the OB and the object were individually placed in each chamber. The OB was allowed to explore freely, while the object was controlled remotely via Bluetooth to move in random directions for 5 min. During conditioning, the OB mouse observed the object moving rapidly and occasionally bumping into the walls for 4 min. b Freezing levels (%) in OBs during Object-threat test were compared to those during OFC with conspecific DMs (right). Freezing levels during the day-2 24-hour memory retrieval, where OBs were placed alone in the same chamber, without the object or DMs (left). Demonstrator: n = 17, Object: n = 11 c Top: Schematic of Object-threat followed by OFC1 for Ca²⁺ imaging. The experimental chamber was changed to a striped context, the same context used in OFC2. Bottom: Three annotated behaviors during the Object-threat test: Threat-freezing, Obj-moving, and Obj-stationary. d Venn diagrams showing proportions of cells encoding Threat-freezing, Obj-moving, and Obj-stationary during the Object-threat session. e Schematic illustrating the procedure for training OFC datasets and test LDA classifiers to decode neuronal activity during the Object-threat session. f LDA decoder performance (n = 6 mice) predicting Threat-freezing, Obj-moving or Obj-stationary during the Object-threat session using OFC1 data (left) and mean decoding accuracy matrix between OFC1 and Object-threat test (right). Source data are provided as a Source Data file. All data are presented as means ± S.E.M.