Evoked Frontal and Parietal Field Potential Signatures of Target Detection and Response Inhibition in Rats Performing an Equiprobable Auditory Go/No-Go Task

Rat behavior. A, Auditory go/no-go discrimination task: rats were presented equiprobably with either a target or distractor tone, followed by a 3-s response window. Licks in response to the target tone (a hit) result in a squirt of water, while failing to lick (a miss) forfeits the opportunity for water reward. Licking in response to the distractor tone is deemed a FA and results in a 10-s penalty period. A CR refers to when the rat successfully refrains from licking after the distractor tone. Licks during the ITI were punished with a 10-s time out. On hit trials, rats were given a 3-s grace period after the response period, during which additional licks would not be penalized. B, Distribution of single-session behavioral performance (percentage correct) of each rat on target (go, black box plots) and distractor (no-go, gray box plots) trials, shown by the median performance (horizontal lines), the typical range (vertical boxes), and putative outlier sessions (plus signs). The right group column plots the distribution of performances for the whole group of nine rats. C, RT histogram of first-lick latencies on hit (black line) and FA (gray line) trials cumulated over 30 go/no-go sessions in nine rats.

Histology. Representative Nissl-stained coronal sections (left panels) showing electrode penetrations in the right hemisphere (next to the 1-mm vertical calibration bar) of dorsomedial frontal (A) and posterior parietal (B) cortex. The right panels show diagrams from the Paxinos and Watson rat brain atlas (Paxinos and Watson, 1997) with their designations for the cortical areas 2.7 mm anterior to bregma (in A; Cg1 is cingulate cortex, M2 and M1 are secondary and primary motor cortex) and 4.16 mm posterior to bregma (in B; PtA is parietal association cortex, RSA is retrosplenial agranular cortex, and CA1 and CA3 are fields of the hippocampus).

Single-trial FPs and single-session frontal and parietal auditory ERPs. A, Single-electrode field potentials (FPs) recorded in frontal (left panel) and parietal (right panel) cortex on three example hit trials. B–D, Single-session hit (black traces) and CR (gray traces) ERPs (in microvolts) recorded in frontal (left panels) and parietal (right panels) cortex of a different rat in each row. The ERPs shown are averaged frontal and parietal FPs referenced to the onset of the target or distractor tone at t = 0 on the horizontal axis. Upwards and downwards triangles label the approximate latencies and amplitudes of the N1, P2, N2, P3E, and P3L peaks identified by our automated analysis.

Comparison of hit and CR trial grand average ERPs and component amplitudes. Grand average ERPs and peak amplitudes from frontal and parietal cortex are shown in the upper and lower panels, respectively. FPs from the frontal or parietal cortex were averaged to produce ERPs (left panels) for hit (black traces and black box plots) and CR trials (gray traces and gray box plots). The ERPs were referenced to the onset of the tone. We used the peak amplitudes of the session ERPs to compute the component amplitudes for each session whose distribution is shown by the box plots in the right panels. Horizontal lines show median peak amplitude, white xs denote the mean amplitude, and plus signs denote putative outliers outside 1.5× the interquartile range. In calculating the grand average ERPs, the single-session ERPs were weighted according to each session’s number of trials to give greater weight to the more statistically reliable (i.e., higher-n) ERPs. The asterisk indicates a significant target detection-related amplification of the P2 peak on hits compared to CRs (Tukey’s HSD, p = 0.006)^f.

Hit versus FA ERPs and component amplitudes. As in Figure 4, Grand average ERPs and peak amplitudes from frontal and parietal cortex are shown in the upper and lower panels, respectively. FPs from the frontal or parietal cortex were averaged to produce ERPs (left panels) for hit (black traces and black box plots) and FA trials (gray traces and gray box plots). The ERPs were referenced to the onset of the tone. We used the peak amplitudes of the session ERPs to compute the component amplitudes for each session whose distribution is shown by the box plots in the right panels. Horizontal lines show median peak amplitude, white xs denote the mean amplitude, and plus signs denote putative outliers outside 1.5× the interquartile range. In calculating the grand average ERPs, the single-session ERPs were weighted according to each session’s number of trials to give greater weight to the more statistically reliable (i.e., higher-n) ERPs. The asterisk indicates a significant amplification of the P2 peak on hits compared to FAs (Tukey’s HSD, p = 0.0009)^g.

Single-session ERP-peak amplitudes by trial type. Each row presents box plots of peak amplitudes (in microvolts) extracted from single-session ERPs for each trial type (hits, misses, CRs, and FAs). Peak amplitudes from frontal and parietal ERPs are shown in the left and right panels, respectively. As in previous box plots horizontal lines show median peak amplitude, white xs denote the mean amplitude, and plus signs denote putative outliers outside 1.5× the interquartile range.

ERP responses to equiprobable passive presentation of lower-pitched and higher-pitched tones equivalent to the distractor and target tones, respectively, in the active go/no-go task. A, B, ERPs and peak amplitudes for two different rats. Upper panels depict frontal and parietal ERPs recorded in response to the lower pitch (black traces) or higher pitch (gray traces), while lower panels show box plots (with conventions as in previous figures) of peak amplitudes for lower pitch (black) or higher pitch (gray) ERPs. The lower pitch (distractor in the active context) tends to elicit a larger P2 peak than the higher pitch; this difference was significant in the pooled data (paired t test, df = 9, p = 0.04)ⁱ. In the passive context, rats were not water restricted and had no access to a lick tube.

Comparison of CR and miss trial ERPs and component amplitudes. As in Figure 4, Grand average ERPs and peak amplitudes from frontal and parietal cortex are shown in the upper and lower panels, respectively. FPs from the frontal or parietal cortex were averaged to produce ERPs (left panels) for CR (black traces and black box plots) and miss trials (gray traces and gray box plots). The ERPs were referenced to the onset of the tone. We used the peak amplitudes of the session ERPs to compute the component amplitudes for each session whose distribution is shown by the box plots in the right panels. Horizontal lines show median peak amplitude, white xs denote the mean amplitude, and plus signs indicate putative outliers outside 1.5× the interquartile range. In calculating the grand average ERPs, the single-session ERPs were weighted according to each session’s number of trials to give greater weight to the more statistically reliable (i.e., higher-n) ERPs. Asterisks indicate significantly greater peak amplitude on miss trials compared to CRs at the N2 (Tukey’s HSD, p = 0.003)^j, P3E (p = 4 × 10⁻⁶)^l, and P3L (p = 0.04)^m peaks.

Distribution of ERP peak latencies on hit (black crosses) and CR (gray crosses) trials. Each cross marks the latency of the respective peak (labeled on the vertical axis) in the ERP from one of 30 recording sessions (from nine rats) in our dataset. The N1 peak was not identified in every session and is omitted from the figure.

Average Hit and CR spectrograms from frontal (left panels) and parietal (right panels) electrodes, aligned to tone onset at t = 0. Hit and CR spectrograms are shown in rows A, B, respectively. Row C shows the difference between average H and CR spectrograms (hit minus CR), such that the light blue and yellow time-frequency bins indicate greater hit power than CR power, and dark blue bins indicate greater CR power than hit power. The color axes in rows A–C show power transformed to arbitrary logarithmic units to allow smaller-amplitude features at higher frequencies to be visible. Row D shows the significance of a two-tailed t test comparing spectral power on hits and CR trials at each time-frequency bin, based on the variability of the trial-averaged power across sessions (n = 30). The statistical significance of the hit-CR difference at each time-frequency bin is plotted as log10(1/p) such that differences from dark blue on the color scale indicates significant differences above the 95% confidence level, while light blue (color scale = 2) and yellow (color = 3) indicate significance with 99% and 99.9% confidence, respectively [i.e., since log10(1/0.001) = 3; not corrected for multiple comparisons; the t tests were based on raw trial-averaged spectral powers in units of square microvolts per Hz, not the logarithmically transformed values displayed in rows A–C].

Average Miss and CR spectrograms from frontal (left panels) and parietal (right panels) electrodes, aligned to tone onset at t = 0. Miss and CR spectrograms are shown in rows A, B, respectively. Row C, shows the difference between average miss and CR spectrograms (miss minus CR), such that the light blue and yellow time-frequency bins indicate greater miss power than CR power, and dark blue bins indicate greater CR power than miss power. The color axes in rows A–C show power transformed to arbitrary logarithmic units to allow smaller-amplitude features at higher frequencies to be visible. Row D shows the significance of a two-tailed t test comparing spectral power on hits and CR trials at each time-frequency bin, based on the variability of the trial-averaged power across sessions (n = 30). The statistical significance of the miss CR difference at each time-frequency bin is plotted as log10(1/p) such that differences from dark blue on the color scale indicates significant differences above the 95% confidence level, while light blue (color-scale = 2) and yellow (color = 3) indicate significance with 99% and 99.9% confidence, respectively [i.e., since log10(1/0.001) = 3; not corrected for multiple comparisons; the t tests were based on raw trial-averaged spectral powers in units of square microvolts per Hz, not the logarithmically transformed values displayed in rows A–C].

Average hit and miss spectrograms from frontal (left panels) and parietal (right panels) electrodes, aligned to tone onset at t = 0. Hit and miss spectrograms are shown in rows A, B, respectively. Row C shows the difference between average hit and miss spectrograms (hit minus miss), such that the light blue and yellow time-frequency bins indicate greater hit power than miss power, and dark blue bins indicate greater miss power than hit power. The color axes in rows A–C show power transformed to arbitrary logarithmic units to allow smaller-amplitude features at higher frequencies to be visible. Row D, shows the significance of a two-tailed t test comparing spectral power on hit and miss trials at each time-frequency bin, based on the variability of the trial-averaged power across sessions (n = 30). The statistical significance of the hit-miss difference at each time-frequency bin is plotted as log10(1/p) such that differences from dark blue on the color scale indicates significant differences above the 95% confidence level, while light blue (color scale = 2) and yellow (color = 3) indicate significance with 99% and 99.9% confidence, respectively [i.e., since log10(1/0.001) = 3; not corrected for multiple comparisons; the t tests were based on raw trial-averaged spectral powers in units of square microvolts per Hz, not the logarithmically transformed values displayed in rows A–C].

Average Hit and CR coheregrams showing frontal-parietal coherence spectra as a function of time relative to tone onset at t = 0. Coheregrams were averaged across trials and interarea electrode pairs, then across sessions. A, Average hit frontal-parietal coheregram. B, Average CR frontal-parietal coheregram. The color axes in A, B depict dimensionless coherence values that can in principle range between zero and one. C, Difference between hit and CR coheregrams (hit minus CR). D, Time-frequency bins where the hit-CR coherence difference is significant at the 99% CL or above are plotted as log10(1/p) where p is the p value resulting from a two-tailed t test based on the variability across sessions.