Fig. 2: BFPVNs of the HDB show strong responses to punishment.
From: Parvalbumin-expressing basal forebrain neurons mediate learning from negative experience
a Raster plots and PETHs of the spike times aligned to cue onset of an example BFPVN. Trials are partitioned by the cue types (Cue 1 predicting likely reward / unlikely punishment, top, green; Cue 2 predicting likely punishment / unlikely reward, middle, purple). b Average PETH of BFPVN activity aligned to cue onset (activation, n = 12, solid line; inhibition, n = 2, dashed line; see inset for distribution of responses; errorshade, SEM). c Example PETH of the same BFPVN as in (a) aligned to reward delivery. Trials are partitioned into expected (following Cue 1) and surprising (following Cue 2) reward trials. d Average PETH of BFPVN activity aligned to reward delivery (activation, n = 14, solid line; inhibition, n = 3, dashed line; see inset for distribution of responses; errorshade, SEM). e Example PETH of the same BFPVN as in (a) aligned to punishment delivery. Trials are partitioned to surprising (following Cue 1) and expected (following Cue 2) punishment trials. f Average PETH of BFPVN activity aligned to punishment delivery (activation, n = 27, solid line; inhibition, n = 2, dashed line; see inset for distribution of responses; errorshade, SEM). g Spike raster plots and PETH aligned to punishment delivery for an example BFPVN. Top, trials in the first half of the session; middle, trials in the second half of the session; bottom, PETH. h Average PETH of BFPVN activity aligned to punishment delivery (n = 27 punishment-activated BFPVNs). Purple, first half of the session; mustard, second half of the session (errorshade, SEM). i Left, difference in peak punishment response between the first and second half of the session (**p < 0.01, p = 0.00458, two-sided Wilcoxon signed-rank rest; n = 27 punishment-activated BFPVNs). Right, difference in baseline firing rate between the first and second half of the session (***p < 0.001, p = 0.00071, two-sided Wilcoxon signed-rank test; n = 27 punishment-activated BFPVNs). Boxes and whiskers show median, interquartile range and non-outlier range. Source data are provided as a Source Data file.
