Skip to main content
. 2021 Jul 21;12:4448. doi: 10.1038/s41467-021-24456-3

Fig. 4. Characteristic timescale of the stimulus-driven population responses and of the intrinsic neuronal activity.

Fig. 4

a Correlation between pairs of population response vectors (as those highlighted by the yellow frames in Fig. 3c) as a function of their temporal lag for two example movies (solid, colored lines). These curves were fitted with either an exponential decay or a damped oscillation function (fits are shown as black dashed lines). The time constants of these fits were taken as the characteristic timescales of the population responses. The solid gray lines show the correlation functions of the corresponding movies, reported from Fig. 3b for comparison. b Timescales of the population responses as a function of the timescales of the corresponding movie stimuli (colored markers). Each colored line is the linear fit prediction of the relationship between such timescales for a given area (same color-code as in the key; note that the lines for LM and LL are partially overlapping). The gray line is the linear fit prediction obtained by pooling together the data of the three extrastriate areas (i.e., LM, LI, and LL). Error bars: standard error of the intercept of the linear fits for V1 and the pooled extrastriate areas from regression analysis (see main text), N = 36 independent combinations of movies and areas. *p = 0.015, two-tailed t-test, no multiple test correction (see main text and “Methods” for details). c, d Same as a and b, but for the intrinsic timescales of neuronal activity (see main text for details). ***p = 5e − 7, 1e − 13, 2e − 14, respectively, for LM, LI, LL. Source data are provided as Source Data files.