Frequency-specific medial septal nucleus deep brain stimulation improves spatial memory in MK-801-treated male rats

Abstract

Background: Few treatments exist for the cognitive symptoms of schizophrenia. Pharmacological agents resulting in glutamate N-methyl-d-aspartate (NMDA) receptor hypofunction, such as MK-801, mimic many of these symptoms and disrupt neural activity. Recent evidence suggests that deep brain stimulation (DBS) of the medial septal nucleus (MSN) can modulate medial prefrontal cortex (mPFC) and hippocampal activity and improve spatial memory.

Objective: Here, we examine the effects of acute MK-801 administration on oscillatory activity within the septohippocampal circuit and behavior. We also evaluate the potential for MSN stimulation to improve cognitive behavioral measures following MK-801 administration.

Methods: 59 Sprague Dawley male rats received either acute intraperitoneal (IP) saline vehicle injections or MK-801 (0.1 mg/kg). Theta (5-12 Hz), low gamma (30-50 Hz) and high frequency oscillatory (HFO) power were analyzed in the mPFC, MSN, thalamus and hippocampus. Rats underwent MSN theta (7.7 Hz), gamma (100 Hz) or no stimulation during behavioral tasks (Novel object recognition (NOR), elevated plus maze, Barnes maze (BM)).

Results: Injection of MK-801 resulted in frequency-specific changes in oscillatory activity, decreasing theta while increasing HFO power. Theta, but not gamma, stimulation enhanced the anxiolytic effects of MK-801 on the elevated plus maze. While MK-801 treated rats exhibited spatial memory deficits on the Barnes maze, those that also received MSN theta, but not gamma, stimulation found the escape hole sooner.

Conclusions: These findings demonstrate that acute MK-801 administration leads to altered neural activity in the septohippocampal circuit and impaired spatial memory. Further, these findings suggest that MSN theta-frequency stimulation improves specific spatial memory deficits and may be a possible treatment for cognitive impairments caused by NMDA hypofunction.

Keywords: Deep brain stimulation; Gamma; Hippocampus; MK-801; Medial septal nucleus; Spatial working memory; theta.

Fig. 1.

Atlas schematic and electrode localization of the MSN-hippocampal-mPFC-thalamic circuit. Representative coronal sections (from anterior to posterior) of the mPFC, MSN, thalamus and hippocampus demonstrate the location of electrode tips. Arrows demonstrate canonical directionality of neural transmission. Rats with stimulating electrodes not within the MSN or recording electrodes not within the mPFC, thalamus or hippocampus were excluded from the study. (Figures E-H modified from the atlas of “The rat brain in stereotaxic coordinates, 5^th Edition” by Paxinos and Watson).

Fig. 2.

Experimental Design. This graphical representation demonstrates the timeline of the experiments. PSD: post-surgery day, LFP: local field potential, IP: intraperitoneal.

Fig. 3.

MK-801 has frequency-specific effects across the septo-hippocampal circuit. Power was compared before and after injection in Saline and MK-801 treated animals. Power spectral densities from the hippocampus (A) before injection and (B) after injection for both groups are shown. Lines indicate mean +/− SEM. (C) The relative change was measured by taking the difference in the power spectral density from the post-injection time period from the pre-injection time period and dividing by the total power of the two time periods together, yielding a normalized difference measure between −1 and 1. Lines indicate mean +/− SEM. This was repeated for the mPFC (D—F), the MSN (G-I), and the Thalamus (J-L). (M-P) To determine the effect of drug, frequency and region on normalized power change, a muti-factorial ANOVA was performed with drug as a between-groups variable and frequency and region as within-groups variables. There was a significant drug × frequency interaction condition (F (1.5,24.7) = 5.84, p = 0.013). This is denoted with a dashed line to indicate this effect was across regions as there was no significant main effect of region (F (1.6,26.7) = 0.65, p = 0.504) or drug × frequency × region interaction (F(3.0, 47.3) = 0.30, p = 0.82).

Fig. 4.

Novel object recognition and elevated plus maze results. (A) ANOVA revealed a significant main effect of stimulation type (F(2,38) = 3.67, p = 0.035) but not of drug (F(1,38) = 0.06, p = 0.807) nor a drug × stimulation type interaction (F(2,38) = 2.97, p = 0.063). Post-hoc pairwise t-tests revealed that saline-gamma stimulated rats spent significantly more time around the novel object than did saline-no stimulation rats (p = 0.046, Bonferroni-Holm corrected) and saline −theta rats (p = 0.046, Bonferroni-Holm corrected). (B) Evaluation of open arm entries on the elevated plus maze using a Kruskel wallis test showed a significant effect of group (X(5) = 19.85, p = 0.001) and post hoc Dunn tests showed a significant difference between MK-801- theta and saline-gamma (p = 0.022, Bonferroni-Holm corrected) and MK801-theta compared to saline-theta (p = 0.002, Bonferroni-Holm corrected). *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 5.

Barnes maze results. (A) There was a significant main effect of stimulation type (F2,50) = 4.55, p = 0.015) and of day (F(2,100) = 28.84, p < 0.0001), and a significant stimulation × drug × day interaction (F(4,100) = 3.23, p = 0.016). Post hoc comparisons revealed that for MK801 treated male rats, theta-stimulated rats had significantly shorter latencies to find the escape hole than both gamma stimulated rats (p = 0.0008, Bonferroni-Holm corrected) and nonstimulated rats (p = 0.014, Bonferroni-Holm corrected). There was no significant difference, however, between gamma stimulated and non-stimulated rats (p = 0.126, Bonferroni-Holm correct) Within saline treated rats, there was no significant difference between non-stimulated rats and theta-stimulated rats (p = 0.261), however, gamma stimulated rats took significantly longer to find the escape hole than nonstimulated rats (p = 0.012, Bonferroni-Holm adjusted). (B) MK-801 treated rats ran significantly faster than saline treated rats (F(1,30) = 14.33, p = 0.0007) but there as no effect of stimulation type (F (2,30) = 0.47, p = 0.626) nor a stimulation × drug interaction (F(2,30) = 0.60, p = 0.556). (C) MK-801-treated rats made a greater number of errors than saline treated rats (F(1,36) = 0.49, p = 0.615), however there was no main effect of stimulation type (F(2,36) = 0.49, p = 0.615) nor a significant stimulation type × drug interaction (F(2,36) = 2.30, p = 0.115). *p < 0.05, **p < 0.01, ***p < 0.001.

Fig. 6.

Barnes maze search strategy. There is a significant difference in search strategies between saline control and MK-801 rats (χ2(df = 2) = 10.23, p = 0.006). (A) Saline control rats utilized spatial or peripheral strategies (51.2% of trials), while (B) MK-801 rats utilized these strategies in only 26.2% of trials. (C) In saline inject rats, MSN theta stimulation led to a more random search strategy in non-stimulated saline rats (χ2(df = 2) = 7.2303, p = 0.027). (D) MSN theta stimulation of MK-801 rats led to a significant improvement in search strategy, as there was an increased use of the spatial search strategies (37.5% of trials) compared to nonstimulated MK-801 rats (χ2(df = 2) = 6.963), p = 0.031) and no difference compared to non-stimulated saline rats (p = 0.657). (E) Saline rats with MSN gamma stimulation utilized more random search strategies throughout the Barnes maze compared to saline non-stimulated rats (χ2(df = 2) = 9.568, p = 0.008). (F) Similarly, MK-801-injected rats with MSN gamma stimulation had different search strategies compared to saline controls rats (χ2(df = 2) = 6.638, p = 0.036) and did not differ from non-stimulated MK-801 rats (p = 0.553).