The prelimbic cortex is critical for context-dependent fear expression

Abstract

The ability to regulate emotional responses in various circumstances would provide adaptive advantages for an individual. Using a context-dependent fear discrimination (CDFD) task in which the tone conditioned stimulus (CS) is paired with the footshock unconditioned stimulus (US) in one context but presented alone in another context, we investigated the role of the prelimbic (PL) cortex in contextual modulation of the conditioned fear response. After 3 days of CDFD training, rats froze more to the CS presented in the fearful than in the safe context. Following bilateral lesions of the PL, rats showed similar levels of freezing to the CS in both contexts, in contrast to the sham-lesioned control animals. The lesions did not impair the rats' ability to discriminate contexts per se, as indicated by intact differential responses in a separate experiment which employed a simple context discrimination task. Consistent with the lesion data, single-unit recordings from the PL showed that the majority of CS-responsive neurons fired at a higher rate in the fearful context than in the safe context, paralleling the behavioral discrimination. Taken together, the current results suggest that the PL is involved in selective expression of conditioned fear to an explicit (tone) cue that is fully dependent on contextual information.

Keywords: amygdala; context; fear discrimination; hippocampus; prelimbic cortex.

Figure 1

Experimental procedure for context-dependent fear discrimination (CDFD). Two different contexts, denoted as A and B, were used for training. During training, rats were given two alternating sessions on the same day, separated by 1 h: one in Context A and the other in Context B. Training lasted for 3 days. In Context A, three parings of the CS and US were presented. In Context B, they were given ten CS-only trials. Training lasted for 3 days. Twenty-four hours after the final training session, the PL lesion and SHAM group received surgery for electrolytic lesion or sham lesion, respectively. After 7 days of recovery, the rats were given two separate test sessions in Context A and B. An additional testing session, in a new context (Context C), was given to determine the magnitude of the CR unaffected by the conditioning contexts. For the recording experiment, the same training procedure was used for 4 days and the recording session was given only 24 h after the last training session. Testing in Context C was omitted.

Figure 2

Histological verification of the PL lesion and acquisition of CDFD. (A) High-resolution scan of a cresyl-violet-stained coronal section shows a representative PL lesion (top), and the reconstruction shows the extent of the electrolytic lesion (bottom). The gray shading depicts the largest and the black depicts the smallest lesion on the matching coronal sections from an atlas (Paxinos and Watson, 1998). (B) During training, all animals gradually acquired differential responses to the CS in Context A and B. (C) After surgery, the SHAM group showed significantly more freezing to the CS in Context A or C than in Context B. However, the PL lesion group showed a similar level of freezing in all three contexts. ^* and ^***denote p < 0.05 and p < 0.001, respectively, compared to the SHAM group in Context B; ^#denotes p < 0.05 compared to the SHAM group in Context A.

Figure 3

Histological verification of the PL lesion and simple context discrimination. (A) A high-resolution scan of a cresyl-violet-stained coronal section shows a representative PL lesion (top), and the reconstruction shows the extent of the electrolytic lesion (bottom). The gray shading depicts the largest and the black depicts the smallest lesion on the matching coronal sections (Paxinos and Watson, 1998). (B) During training, all animals gradually acquired fear responses to Context A but not to Context B. (C) After the surgery, both SHAM and PL lesion groups continued to show differential response to the two contexts as indicated by a higher level of freezing to Context A than that to Context B. ^*denotes p < 0.05.

Figure 4

Recorded locations and acquisition of CDFD. (A) A high-resolution scan of a cresyl-violet-stained coronal section shows a representative electrode placement in the PL marked by a small lesion (left), and the reconstruction of electrode placements from all subjects (right) shows that the recording locations were confined within the PL. (B) Robust CDFD was developed following 4 days of training as all animals exhibited significantly greater freezing to the CS in Context A than in Context B on days 3 and 4. (C) On the test session, they also showed significantly more freezing to the CS in Context A than in Context B. ^* and ^**denote p < 0.05 and p < 0.01, respectively.

Figure 5

Recorded activities of short-latency CS-responsive PL units during CDFD. (A) Waveforms of two representative short-latency CS-responsive units, their raster plots, and the peristimulus time histograms (PSTHs) are presented. Both units showed increased firing rates to the CS in Context A than in Context B. (B) Overall activity of short-latency CS-responsive PL units during the CS presentation was greater in Context A than in Context B. (C) The mean firing rate during the initial 150-ms period after the CS onset was significantly higher in Context A than in Context B. ^**denotes p < 0.01.

Figure 6

Representative and average activity of PFUs during CDFD. (A) Two representative waveforms of CS-responsive neurons and their raster plots and PSTHs are presented. Both units showed an increased firing rate to the CS in Context A than in Context B. Note that the first representative unit is the same as shown in Figure 5A. (B) The population firing pattern of all CS-responsive units in Context A and Context B are illustrated. The spike firing of PFUs to the CS was higher in Context A than in Context B. (C) The mean firing rate during the initial 10-s period after the CS onset was significantly higher in Context A than in Context B. ^**denotes p < 0.01.

Figure 7

Modulation of activity in PFUs. The direction and amplitude of the modulation are expressed as paired t-values for each unit, which were computed from Z-score differences between Context A and B. The t-values were arranged in the order from the most negative to most positive values. A positive value indicates more activity in Context A. The dark bars represent statistically significant differences in firing rate between the two contexts and white bars represent non-significant differences.

Figure A1

Effects of PL lesions on general activity and sensory gating. There were no differences between groups in total distance (A), the ratio of the duration in marginal and central area (B), and rearing (C). In the PPI test, the ratio of inhibition was not different between groups (D).

Figure A2

Mean percentage of freezing during the pre-tone period before the presentation of the first CS. (A,B) In the lesion experiment, both SHAM and PL lesion groups showed a similar level of pre-tone freezing across the 3 days of training (A). However, on the test session (B), the SHAM group developed a pre-tone response that discriminates Context A from other contexts, as opposed to the PL lesion group which continued to show non-differential responses in all three contexts. (C) In the recording experiment, the rats displayed similar levels of pre-tone freezing in Context A and B across the first three training sessions and testing, but not on training day 4 in which they showed differential responses. ^* and ^***denote p < 0.05 and p < 0.001, respectively (RM ANOVA followed by Fisher's least significant difference test).