Fig. 5: SPA and SPI are quantized in units of afferent neocortical UDS cycles.

a In the model, the efferent SPA (blue) span a continuous range of durations, but they are quantized (green boxes, 1, 2, 3) in the units of afferent UDS cycles (gray). b Using simulated UDS, all efferent Up states were aligned to the efferent Down-Up transition and ordered with increasing duration. Each row represents a single efferent Down-Up transition. c A second matrix was constructed using the same time points, but using the afferent network activity. The examples (1-3) in (a) correspond to the row numbers in (b/c). Efferent SPA can last 8 s, spanning several afferent UDS cycles. Same color axis shows amplitude (in z-score) for all the panels. d In vivo data from an MECIII neuron, matched to the simulation used in (a–c), validates the model by showing efferent MECIII Up states (blue) that last integer multiples (i-iii) of afferent neocortical LFP UDS cycles (gray). e Similar to (a–d), but for SPI, showing that efferent Down state durations show a continuous range, but are quantized (red boxes) in the units of afferent UDS cycles, in the model and (f) in vivo. g With time measured in units of the varying afferent UDS cycles, the model (inset) predicted that the distribution of both Up and Down state durations should be multimodal, with peaks at the half integers (reflecting state transitions after an Up/Down state). In vivo data combined from all experiments showed the predicted multimodality and quantization when ECIII Up/Down state durations were measured w.r.t. variable neocortical UDS cycle lengths. The model prediction and the in vivo data matched over three orders of magnitude variation of SPA and SPI probabilities.