This behavioral test set-up was previously established in our lab to measure anxiety caused by traumatic re-experiencing animals (Ritov and Richter-Levin, 2014). It consisted of an annular platform (90 cm diameter; 10 cm width), made out of black plywood, and joined to the UWT plastic tank. The annular platform had two opposite closed quadrants (with walls 35 cm height) and two open quadrants (with borders 5 mm height). For the tests, rats were first habituated to the room for 5 min and then were placed into one of the open quadrants facing a closed part of the apparatus. Rats were allowed to explore the arena for 5 min. During this time, behavior of the rat was tracked, recorded and analyzed by the Etho-Vision system (Noldus Information Technology, Wageningen, Netherlands).

Animals were exposed to JVS at PND 27 and UWT at PND 60. Behavioral tests, including WAZM and EPM, were conducted 2 weeks post UWT to differentiate animals into trauma affected or unaffected. Trauma affected animals were given fluoxetine treatment for a period of 30 days in drinking water. At the end of the treatment, WAZM and EPM test were repeated to classify animals as treatment responders or non-responders. Animals were sacrificed and brain tissues were harvested at the conclusion of the behavioral tests. Western blot analysis was conducted to check expression of α1, α2 subunit of GABA_A receptor and GluN1 and GluN2A subunit of NMDAR in the (i) ventral dentate gyrus (vDG), (ii) ventral CA1 (vCA1), (iii) ventral CA3 (vCA3) (Figure 1).

Based on the performance on WAZM and EPM, 7 behavioral parameters that were reflective of explorative activity and anxiety levels were chosen (Supplementary Figure 1). One standard deviation was used to decide the cut-off value for each behavioral parameter based on the performance of the control animals. Animal whose behavioral parameter value was either lower or higher than the cut-off value was considered affected in that specific parameter. Post-UWT, animals which were found to be affected in at least 4 out of 7 parameters were classified as trauma affected. Similarly post treatment animals were required to exhibit less than 4 affected parameters to be considered treatment responders. The following parameters were used to classify the animals as trauma affected and unaffected or fluoxetine responders and non-responders:

Rats were randomly assigned to one these two groups: controls (n = 16) and JVS + UWT i.e., Trauma exposed (n = 93). IBP 1 was used to classify animals as trauma-affected or trauma-unaffected. The trauma-affected) rats (n = 53) were treated with fluoxetine for 30 days. Post-treatment animals were further classified as treatment responders or non-responders, again based on IBP 1. Brain tissue was harvested and the ventral hippocampus dissected out from a subset of these animals. Western-blot experiment was conducted for responders which showed complete symptom remission (n = 6), non-responders (n = 4) that continued to be severely affected on all behavioral parameters and unaffected control animals (n = 5).

Animals were sacrificed and the brains were snap frozen in dry ice to keep the protein stability intact. On the cryostat, frozen tissue punches were collected (diameter 1 mm, depth 1.5 mm) from vDG, vCA1, and vCA3—7.6 from Bregma, horizontal orientation of the brain.

One milliliter of lysis buffer was prepared for tissue homogenization (HEPES 10 mM, EDTA 2 mM, EGTA 2 mM, DTT 0.5 mM, protease inhibitor 1:100 dilution, phosphatase inhibitor 1:100 dilution, SDS 0.5% and double distilled water). Hundred microliter of this solution was used to homogenize tissue samples. Protein concentration was estimated using Bradford assay.

Ten microgram of protein was loaded on a 10% SDS-polyacrylamide gel for electrophoresis (SDS-PAGE). The proteins from the gel were transferred to nitrocellulose membrane using semi-dry transfer technique. The proteins on the membrane were visualized using ponceau staining. The membrane was cut according to the molecular weight marker for the protein of interest. The nitrocellulose membrane was washed with double distilled water and 1× TBST to remove ponceau staining. The membrane was later blocked with 5% milk in 1× TBST for 2.5 h. After blocking, the membrane was washed three time with 1× TBST. The membranes were incubated overnight at 4°C with rabbit polyclonal to GABA_A α1 1: 5,000 (Abcam—ab33299), rabbit polyclonal to GABA_A α2 1: 1,000 (Abcam—ab72445), rabbit polyclonal to GluN1 1:1,000 (Abcam—ab521717), rabbit polyclonal GluN2A 1:1,000 (Abcam—ab169873) or rabbit polyclonal GAPDH 1:10,000 (Abcam—ab9485). The next day, membranes were washed thrice with 1× TBST and incubated for 1 h with Goat α-rabbit secondary antibody 1: 10,000 (Abcam—ab6721). The membranes were washed thrice and developed using ECL plus substrate for chemiluminiscence. Since the molecular weight of α1 and α2 of GABA_A receptor are the same, two separate gels were run for each sample. The optical density of the signals was measured using Quantity1 analysis software. The optical density for each band was first normalized to background intensity and then with its respective GAPDH signal. The optical density was then normalized to mean density of the control group for each target protein and region.

Animals were subjected to the same protocol of JVS at PND 27 and adulthood UWT at PND 60. Two weeks post UWT, animals were categorized into trauma affected or unaffected based on their behavior in WAZM, OFT, and SRT. Trauma affected animals were administered fluoxetine treatment for 30 days in drinking water. At the end of treatment animals were subjected to WAZM, OFT, SRT, and EPM. The animals were behaviorally profiled based on their behavior in WAZM, EPM, and SRT. At the end of the behavioral battery the animals were taken for in-vivo electrophysiological study to investigate local circuit activity and long term potentiation in the dorsal dentate gyrus (dDG) (Figure 2).

In order to classify animals as trauma affected or unaffected post UWT, behavioral parameters were selected from WAZM, SRT, OFT. The behavioral parameters of WAZM, SRT, EPM were used to classify animals as treatment responders or non-responders. The rationale for replacing OFT with EPM at the post treatment time point was the observation that control animals did not remain curious enough to explore the same arena on their second exposure to OFT. We replaced the two parameters of OFT that measured explorative activity and anxiety-like behavior, i.e., (i) total distance traveled and (ii) time spent in OFT center at the post trauma time point with two similar readouts from EPM at the post-treatment time point namely (i) total distance traveled in EPM and (ii) Anxiety index EPM open arm time. Similar to IBP1, behavioral cut-off values based on one standard deviation from the performance of the control animals were determined for each parameter and for each time point. Animals affected in at least 4 out of 7 parameters were considered trauma affected post trauma and those who showed recovery in at least 4 parameters after fluoxetine treatment were considered treatment responders.

The following parameters were used post-trauma to classify the animals as trauma affected and unaffected.

In addition to the first five parameters mentioned above the following two parameters of EPM were used at the post-treatment time point.

This experimental set of animals was divided into two groups: controls (n = 16) and Trauma exposed (n = 72). IBP 2 was used to classify animals as trauma-affected or trauma-unaffected. Fluoxetine was administered to trauma-affected rats (n = 35) for a duration of 30 days. IBP 2 was used to classify animals as treatment responders or non-responders post-fluoxetine treatment. These animals were later subjected to electrophysiological tests to measure local circuit activity and long-term potentiation in dDG. For measuring local circuit activity, the following number of animals were analyzed: controls (n = 14), responders (n = 12), and non-responders (n = 18). Animals which either died during the protocol, were outliers, or did not have stable baseline activity were excluded from the analysis. For the LTP measurement the data of controls (n = 11), responders (n = 8), and non-responders (n = 13). Animals were excluded from analysis if baseline was found unstable.

Animals were anesthetized (40% urethane, 5% chloral, hydrate in saline, max. 0.5 ml/100 g i.p.) and placed in a stereotaxic apparatus. Body temperature (maintained at 37 ± 0.5°C with Homeothermic Blanket System, Harvard apparatus, Dover, MA, USA) and level of anesthesia were monitored and adjusted if necessary during the whole experiment. The scalp was opened and of small bur holes (1 mm diameter) were drilled. A glass recording electrode (tip diameter 2–5 μm, filled with a 2 M NaCl solution) was inserted into the dorsal DG (coordinates: –4 mm AP, 2.5 mm ML from Bregma and appr. –3.7 mm DV from brain surface) and a bipolar 125 μm stimulating electrode was placed in the perforant path (PP; coordinates: 8.0 mm AP and 4 mm ML from bregma, –3 mm DV from brain surface. The final depth of the electrodes was adjusted to yield maximal excitatory postsynaptic potentials (EPSP), evoked by a single pulse delivered to the ipsilateral PP (0.1 ms duration; rectangular monophasic). The characteristic features of the depth profile of the DG response thereby served as an indicator for correct electrode location. Evoked field potentials were amplified (×1,000), bandpass filtered at 0.1–1,000 Hz (AM-Systems amplifier), digitized at 10 kHz (CED, Cambridge, UK) and stored to disk for off-line analysis using Spike-2 software (version 4.24, CED, Cambridge, UK), where the field EPSP slope and the population spike (PS) amplitude were measured.

After electrode insertion, recordings were allowed to stabilize for 20 min before recording of an input-output curve commenced (stimulation intensities from 0.4 to 3.2 mA, 0.2 ms duration) to determine the baseline stimulation intensity (40% of maximal field potential response evoked). After 30 min of baseline recording (at 0.1 Hz), different protocols were applied to determine local circuit activity:

For each protocol, the inhibition strength was expressed as percentage of the baseline response, for the population spike amplitude and the EPSP slope, respectively.

For measuring LTP, baseline recordings of EPSP were assessed during 20 min, at frequency of 0.1 Hz. LTP was induced by theta burst stimulation (TBS) (Ardi et al., 2014): three sets of 10 trains, each with 10 pulses (100 Hz), were administrated with 200 ms inter-train interval, and 1 min inter-set interval. LTP was recorded for 60 min after TBS.

Individual behavioral profiling-1 revealed that not every animal that was exposed to JVS+UWT was trauma affected. Nevertheless, the majority of trauma-exposed animals developed PTSD-like symptoms (∼65.6%), while around (34.6%) of the animals remained unaffected (Figure 3B). Proportion of trauma affected individuals was found to be significantly higher in the trauma-exposed group than the controls group (Fisher exact test, p < 0.0001). Only 12.5 % animals of the control group were found to be affected (Figure 3A).

The average group effect on behavior parameters indicated that the trauma-affected animals were severely affected in most of the behavioral measures. (i) Total distance traveled in WAZM revealed significant main effect of the groups [H (3) = 67.36, p < 0.0001] as measured by Kruskal–Wallis test. Dunn’s post-hoc comparison further revealed that the trauma-affected animals traveled significantly less distance than both the control (p < 0.0001) and unaffected animals (p < 0.0001) (Figure 3C). (ii) Time spent in the closed arms of the WAZM was compared across the three groups and Kruskal–Wallis test showed significant difference between groups [H (3) = 34.53, p < 0.0001]. Dunn’s post-hoc multiple comparison verified that the trauma affected animals spent more time in the anxiogenic closed arms of WAZM than control (p < 0.0001) and unaffected (p < 0.0001) animals (Figure 3D). (iii) Distance traveled in closed arms of the WAZM was found to be significantly different across the groups as shown by Ordinary-One-Way ANOVA [F_(2,106) = 27.31, p < 0.0001]. Bonferroni post-hoc comparison showed that the affected animals traveled less in the WAZM closed arms than both control (p < 0.0001) and unaffected (p < 0.0001) animals (Figure 3E). Significant difference between the three groups was found in the iv) distance traveled in open arms by the Kruskal–Wallis test [H (3) = 61.24, p < 0.0001]. Dunn’s post-hoc multiple comparison further revealed that distance traveled in WAZM open arms by affected group was less than both control (p < 0.0001) and unaffected animals (p < 0.0001) (Figure 3F). No main effect of group was found by ordinary one-way ANOVA in the (v) total time spent freezing in WAZM across the 3 groups (Figure 3G). In the EPM test, significant main effect of group was found in (vi) total time spent freezing in EPM using ordinary one-way-ANOVA test across the three groups [F_(2,106) = 35.13, p < 0.0001] Bonferroni’s multiple comparison test showed increased freezing of affected animals than control (p < 0.0001) and unaffected (p < 0.0001) animals (Figure 3H). Kruskal–Wallis test showed significant difference between groups in (vii) anxiety index in closed arms of EPM [H (3) = 19.06, p < 0.0001]. Dunn’s post-hoc test revealed affected animals had an increased anxiety index as compared to control (p < 0.05) and unaffected animals (p < 0.001) (Figure 3I). In all the parameters no significant difference was found in the behavior between the control and unaffected animals. This showed that the unaffected animals were behaviorally similar to control animals. The observation that the trauma-affected animals had worse average group effects on behavioral parameters than controls and unaffected animals reinforced the fact that IBP1 effectively categorized affected and unaffected animals post-UWT.

Individual behavioral profiling-1 successfully identified animals that responded to the fluoxetine treatment. Around 49% of the trauma affected animals that received fluoxetine treatment showed recovery in ≥4 out of 7 behavioral parameters (Figure 4B). Fifty-one of the trauma-affected animals did not show recovery and were classified as non-responder (Figure 4B). This finding corroborated human statistical data indicating that only a subset of PTSD patients respond to therapy. IBP-1 also profiled the control animals as affected or unaffected with the percentage of unaffected animals much higher than affected animals (Figure 4A).

The behavior profiling was further supported by the differences observed in the group average effect of responders and non-responders on behavior. The average group effect of the treatment responders on most behavioral parameters was either found to be similar to healthy control animals or better than the non-responders. Kruskal–Wallis test showed significant difference between the three groups in the (i) total distance traveled in WAZM [H (3) = 41.51, p < 0.0001]. Dunn’s post-hoc comparison revealed that the non-responders traveled significantly less distance than both controls (p < 0.0001) and responders (p < 0.00001). The responders and controls traversed similar distance in the WAZM (Figure 4C). Similarly, significant main effect of groups was found in (ii) time spent in closed arms as compared by Kruskal–Wallis test [H (3) = 20.62, p < 0.0001]. Dunn’s post-hoc comparison revealed that non-responders spent significantly more time in the closed arms of the WAZM than controls (p < 0.00001) and responders (p = 0.0151). No significant difference was found in the time spent in closed arms between the controls and responders (Figure 4D). The average group effect was measured for (iii) distance traveled in closed arms of WAZM using Kruskal–Wallis test and was found to be significantly different between the groups [H(3) = 28.11, p < 0.00001]. Dunn’s post-hoc comparison showed that non-responders traversed significantly less distance in the closed arms of the WAZM than controls (p = 0.0019) and responders (p < 0.0001). The controls and responders traversed similar distance in the closed arms and no significant difference was found between them (Figure 4E). Significant group differences were also found in the iv) distance traveled in the open arm of the WAZM as analyzed by Kruskal–Wallis test [H (3) = 27.21, p < 0.0001]. Dunn’s post-hoc comparison revealed that the non-responders traveled significantly less distance in the open arm as compared to controls (p < 0.00001) and responders (p < 0.00001). No significant difference was found between the controls and responders in the distance traveled in the open arms (Figure 4F). Similar trend was observed in the (v) total time spent freezing in the WAZM by the three groups. Significant main effect of groups was found in Kruskal–Wallis test [H (3) = 34.13, p < 0.0001) and Dunn’s post-hoc comparison showed that the controls are behaviorally same as responders but the non-responders spent significantly more time freezing than the controls (p < 0.00001) and responders (p < 0.00001) (Figure 4G). No significant main effect of groups was found in the (vi) total time spent freezing in the EPM between the groups (Figure 4H). However, significant differences was found between the groups in the (vii) anxiety index time in EPM as measured by Kruskal–Wallis test [H (3) = 11.85, p < 0.01]. Dunn’s post-hoc comparison further revealed that the non-responders had significantly higher anxiety index than the controls (p = 0.0117) and responders (p = 0.0102). No significant difference was found in the anxiety index between the controls and responders (Figure 4I). The results clearly showed that behaviorally, treatment-responders were similar to controls, while non-responders continued to perform worse in all behavioral parameters than both controls and responders.

Individual behavioral profiling-2 revealed that 57% of trauma-exposed animals were affected (Figure 6B), compared to 12.5% of controls (Figure 6A). According to Fisher exact test, the percentage of affected animals in the JVS+UWT group was significantly greater than the controls (p = 0.0017). The distribution of affected and unaffected individuals obtained using IBP-2 was compared to that obtained from IBP-2 using Fisher exact test but no statistically significant difference was identified (p = 0.1423).

The average group effect on the 7 behavioral parameters was compared between the trauma affected, unaffected and control animals. Statistically significant difference was found in the (i) distance traveled in the closed arms of WAZM between the groups [H(3) = 18.25, p = 0.0001], as measured by Kruskal–Wallis test. Dunn’s post-hoc multiple comparison showed that the affected individual traveled significantly less distance in closed arms than controls (p = 0.0039) and unaffected animals (p = 0.0006). The controls and unaffected animals traveled similar difference in the closed arms of WAZM and no significant difference was found between the two groups (p > 0.0999) (Figure 6C). (ii) Time spent in the open arms of the WAZM was compared using Kruskal–Wallis test and was found to be significantly different across groups [H(3) = 27.03, p < 0.0001]. Dunn’s post-hoc multiple comparison showed that the affected animals spent significantly less time in the anxiogenic open arms than the controls (p = 0.0011) and unaffected animals (p < 0.00001). The controls and unaffected animals spent similar time in the open arms of the WAZM with no significant statistical difference found between them (p > 0.0999) (Figure 6D). (iii) Distance traveled in the open arms of the WAZM was found to be significantly different between groups when measured using Kruskal–Wallis test [H(3) = 39.43, p < 0.0001). Dunn’s post-hoc test showed that the affected animals traveled less distance in the anxiogenic open arms than controls (p = 0.0001) and unaffected animals (p < 0.00001). No significant difference was found between the controls and unaffected animals (p > 0.9999) and both the groups appeared to travel similar distance in the open arms of WAZM (Figure 6E). Statistically significant difference was found between groups in the (iv) total time spent freezing in the WAZM using ordinary-one way-ANOVA and was found to be significantly different between groups [F_(2,82) = 17.69, p < 0.0001]. Bonferroni’s post-hoc test revealed that the affected individual spent significantly more time freezing in the WAZM than controls (p < 0.00001) and unaffected animals (p < 0.00001). No significant difference was found between the controls and unaffected individual in terms of total time spent freezing in the WAZM (p > 0.9999) (Figure 6F). (v) Total distance covered in the OFT was compared between the three groups using Kruskal–Wallis test and was found to be significantly different [H(3) = 21.22, p < 0.00001]. Dunn’s post-hoc comparison test showed that the affected individuals traveled significantly less distance in the OFT as compared to controls (p = 0.0002) and unaffected animals (p = 0.0018). There was no significant difference in the total distance traveled in OFT between the controls and the unaffected (p = 0.5335) (Figure 6G). (vi) Time spent in the center of OFT was compared between groups and was found to be significantly different using Kruskal–Wallis test [H(3) = 33.98, p < 0.00001]. Dunn’s post-hoc comparison revealed that the affected individuals spent significantly less time in the anxiogenic center of the OFT than controls [p = 0.0003] and unaffected individuals [p < 0.00001]. No statically significant difference was found between the controls and unaffected (p > 0.9999). No significant difference was found between the groups in the (vii) preference index to novel animal in SRT using ordinary-one way-ANOVA [F_{(2, 82)} = 2.210, p = 0.1160] (Figures 6H,I).

Together, the results indicated that the behaviorally profiled affected animals displayed increased anxiety-like behavior. The animals which were profiled as unaffected by IBP-2 were behaviorally similar to the healthy controls giving further validation to the behavioral profiling approach.

Individual behavioral profiling-2 successfully distinguished treatment responders from non-responders after fluoxetine therapy. Thirty-seven percent of the affected population was found to be treatment responders and 63% did not respond to the treatment (Figure 7B). The control population was also screened for any affected individuals, however the percentage of affectedness remained low at 18.5% (Figure 7A). No statistically significant difference was found in the distribution of responders and non-responders between experiment 1 and experiment 2 as compared by Fisher’s test (p = 0.2842).

The average group effect of the responders, non-responders and controls was measured for the 7 behavioral parameters chosen. (i) Distance travelled in the closed arms of WAZM was compared across groups using Kruskal–Wallis test and was found to be significantly different [H (3) = 12.42, p = 0.0020]. Dunn’s post-hoc comparison showed that the non-responders traveled significantly less than controls in the closed arms (p = 0.0013). The mean rank difference between responders and controls was 9.670, and between responders and non-responders was 7.524, although neither difference was statistically significant (p = 0.23367, p = 0.3967 respectively) (Figure 7C). (ii) Time spent in the open arms of the WAZM was found to be significantly different across groups using Kruskal–Wallis test [H (3) = 22.99, p < 0.00001). Dunn’s post-hoc test revealed no significant difference in the time spent in the anxiogenic open arms between the responders and controls (p = 0.8052). However, the non-responders spent significantly less time in the anxiogenic open arms of the WAZM than controls (p < 0.00001) and responders (p = 0.0036) (Figure 7D). iii) Distance travelled in the open arms of the WAZM was found to be significantly different between the groups as seen by Kruskal–Wallis test [H (3) = 29.21, p < 0.00001]. Further post-hoc comparison using Dunn’s test revealed that the non-responders travelled significantly less in the open arms than the controls (p < 0.00001) and the responders (p = 0.0141). No statistically significant difference was found between responders and controls (p = 0.0788) (Figure 7E). iv) Total time spent freezing in the WAZM was compared between the groups using ordinary-one way ANOVA and was found to be significantly different [F_{(2, 48)} = 10.52, p = 0.0002]. Bonferroni’s post-hoc test showed that the non-responders spent significantly more time freezing in the WAZM than the controls (p = 0.0001) (Figure 7F). (v) Total distance traveled in EPM was found to be significantly different between the groups as compared by ordinary-one way- ANOVA [F_{(2, 48)} = 17.23, p < 0.00001]. Bonferroni’s post-hoc test revealed that non-responders traveled significantly less distance in the EPM than controls (p < 0.00001) and responders (p = 0.0001). No statistically significant difference was observed between the responders and the controls (Figure 7G). (vi) Anxiety index in the closed arms of EPM was observed to be significantly different between the groups as measured by sung Kruskal–Wallis test [H(3) = 10.85, p = 0.0044]. Dunn’s post-hoc comparison showed that the non-responders exhibited significantly higher anxiety index than the controls (p = 0.0042). No significant difference was observed between responders and controls (p = 0.9197) or non-responders (p = 0.1622) (Figure 7H). No statistical difference was observed between the responders and controls (p = 0.0937) or non-responders (p = 0.1390). Kruskal–Wallis test revealed no significant difference between groups in the vii) preference index for novel animal in SRT (Figure 7I). The findings revealed that responders recovered in most behavioral parameters and were either behaviorally equivalent to controls or better than non-responders. This demonstrated that IBP-2 was effective in detecting responders within the fluoxetine-treated group.

The feed-forward inhibition in the dDG was measured using FDI protocol but the EPSP slope was not found to be significant between the group as measured by Kruskal–Wallis test [H(3) = 2.609, p = 0.2713] (Figure 8A). No statistically significant difference was found in the FDI population spike (PS) amplitude between groups as revealed by ordinary-one way-ANOVA [F_{(2, 41)} = 0.3740] (Figure 8A). However increased feed-back inhibition was observed in the responders when the local circuit activity was measured using PPI protocol. PS amplitude after 15ms ISI of PPI protocol was significantly different between groups as seen by using Kruskal–Wallis test [H (3) = 10.27, p = 0.0059]. Dunn’s multiple comparison revealed that the responders had increased local-circuit inhibition after 15ms protocol in PPI as evinced by significantly low PS amplitude as compared to controls (p = 0.0066) and non-responders (p = 0.0373). No statistically significant difference was observed between controls and non-responders (p > 0.9999) (Figure 8B). PS amplitude induced by PPI 30 ms protocol was significantly different between groups [H (3) = 9.366, p = 0.0093]. Dunn’s multiple comparison showed that the responders has significantly reduced population spike induction than controls (p = 0.0268) and non-responders (p = 0.0161). Majority of control and non-responder animals showed facilitation in the local-circuit after 30ms ISI between pulses unlike most of the responders that exhibited feed-back inhibition. No statistically significant difference was observed between controls and non-responders (p > 0.9999). The EPSP slope induced by 15ms (Figure 8B) and 30 ms (Figure 8C) protocol of PPI were found to be non-significant between groups as measured by Kruskal–Wallis test [H(3) = 4.230, p = 0.1207] and [H(3) = 3.792, p = 0.1502], respectively. No significant difference was observed between groups in EPSP slope or PS amplitude induced by PPI 80ms protocol as measured by ordinary-one way-ANOVA [F_(2,41) = 0.7357, p = 0.4857] and [F_(2,41) = 0.7695, p = 0.4700], respectively (Figure 8D).

No differences in EPSP or PS amplitude was discovered after TBS in LTP induction protocol as well (Supplementary Figure 2). Together, increased feed-back inhibition in the dDG local circuit was displayed only by the responders and was restricted to decreased PS amplitude in PPI 15 and PPI 30 protocol. This not only highlighted electrophysiological difference between the responders and the others but also highlighted a unique functional adaptive response of the responders to fluoxetine.