Modulation of Theta-Band Local Field Potential Oscillations Across Brain Networks With Central Thalamic Deep Brain Stimulation to Enhance Spatial Working Memory

Abstract

Deep brain stimulation (DBS) is a well-established technique for the treatment of movement and psychiatric disorders through the modulation of neural oscillatory activity and synaptic plasticity. The central thalamus (CT) has been indicated as a potential target for stimulation to enhance memory. However, the mechanisms underlying local field potential (LFP) oscillations and memory enhancement by CT-DBS remain unknown. In this study, we used CT-DBS to investigate the mechanisms underlying the changes in oscillatory communication between the CT and hippocampus, both of which are involved in spatial working memory. Local field potentials (LFPs) were recorded from microelectrode array implanted in the CT, dentate gyrus, cornu ammonis (CA) region 1, and CA region 3. Functional connectivity (FC) strength was assessed by LFP-LFP coherence calculations for these brain regions. In addition, a T-maze behavioral task using a rat model was performed to assess the performance of spatial working memory. In DBS group, our results revealed that theta oscillations significantly increased in the CT and hippocampus compared with that in sham controls. As indicated by coherence, the FC between the CT and hippocampus significantly increased in the theta band after CT-DBS. Moreover, Western blotting showed that the protein expressions of the dopamine D1 and α4-nicotinic acetylcholine receptors were enhanced, whereas that of the dopamine D2 receptor decreased in the DBS group. In conclusion, the use of CT-DBS resulted in elevated theta oscillations, increased FC between the CT and hippocampus, and altered synaptic plasticity in the hippocampus, suggesting that CT-DBS is an effective approach for improving spatial working memory.

Keywords: central thalamus; deep brain stimulation; hippocampal theta oscillation; spatial working memory; synaptic plasticity.

FIGURE 1

Experimental procedure of CT-DBS evaluation. Thirty rats were divided into sham controls (N = 10), DBS group (N = 10), and DBS wo. T-maze group (N = 10). After a 7-day surgical recovery, implanted rats in sham controls and DBS group received CT-DBS (or sham stimulation) and underwent T-maze behavioral task training and LFP recording. For confirmation of CT-DBS-induced dynamic changes in LFP oscillatory activity in different hippocampal regions, to exclude confounding effects of T-maze behavioral training, we added a group of DBS without T-maze task (DBS wo. T-maze group, marked by dashed box). LFP recording (green bar): LFP recording was performed twice, i.e., on the 8th day for baseline and before Western blotting on the 16th day. Sham CT-DBS (blue bar): rats were placed in a plastic cage without 30-min CT-DBS and then trained for the T-maze behavioral task for 7 days. For CT-DBS (yellow bar): rats were placed in the same plastic cage with 30-min CT-DBS and then trained for the T-maze behavioral task for 7 days. Rats in sham controls and DBS group were sacrificed for protein level analysis by Western blotting on the 17^th day (black bar).

FIGURE 2

The implantation sites of the microelectrode array was confirmed by Nissl staining. A 16-channel stainless microwire electrode array was used to perform CT-DBS and multi-site LFP recordings. (A) The representative coronal brain slice shows two small electrolytic lesions made around the tip of the electrode by passing 50 μA DC for 30 s in the bilateral CT (AP, −2.7 mm; ML, ±1.6 mm; and DV, 5.5 mm) marked with two star-symbols (^∗). (B) Three electrolytic lesions marked with three star-symbols (^∗): left hippocampal DG (AP, −4.0 mm; ML, −2.0 mm; and DV, 3.6 mm), CA1 (AP, −4.0 mm; ML, −2.4 mm; and DV, 1.9 mm), and CA3 (AP, −4.0 mm; ML, −3.0 mm; and DV, 1.9 mm).

FIGURE 3

Comparsion of behavioral performances in the T-maze behavioral task between the groups following CT-DBS (or sham stimulation). (A) Curves of latency time (s) to reach the correct T-maze goal arm was plotted against the sessions, i.e., one session/day and five trials/session. There were significantly shorter learning periods found in the DBS group from the 13^th to 15^th day compared with those in the sham controls during T-maze behavioral task training. (B) SWMI showed the percentage of the mean correct ratio to reach the correct T-maze goal arm against the sessions as (A). SWMI also was found to significantly increase from the 12^th to 15^th day in the DBS group. ^∗∗, and ^∗∗∗ indicate significant differences in terms of latency time and SWMI with p < 0.01 and p < 0.001, respectively, analyzed by Wilcoxon two-sample t-tests (mean ± SEM).

FIGURE 4

Comparsion of LFP oscillatory changes before (baseline) and after CT-DBS (or sham stimulation) in both groups. The normalized percentage of the LFP PSD was calculated as the ratio of the original PSD before CT-DBS or sham stimulation (baseline) to that after CT-DBS or sham stimulation in the CT, DG, CA1, and CA3. (A) In the sham controls, there were no significant differences in terms of delta, theta, alpha and beta bands at each site. (B) In the DBS group, LFP PSD showed significant enhancements for the theta and alpha bands in the CT, DG, CA1, and CA3 following CT-DBS. ^∗, ^∗∗, and ^∗∗∗ indicate significant differences in terms of PSD with p < 0.05, p < 0.01, and p < 0.001, respectively, analyzed by Wilcoxon signed-rank tests (mean ± SEM).

FIGURE 5

Statistical comparison of FC changes before (baseline) and after CT-DBS (or sham stimulation). FC analyses were used to estimate the normalized coherence between the pairs of brain regions at LFP recording sites (CT, DG, CA1, and CA3) in the two group. (A) The strengths of theta- and alpha-band coherences in FC showed no significant differences before and after sham stimulation in the sham controls. (B) In DBS group, FC strengths between the brain region pairs CT–DG, CT–CA1, CT–CA3, DG–CA1, DG–CA3, and CA1–CA3 significantly increased with theta-band coherences following 7-day CT-DBS. Further, the FC strengths between the brain region pairs CT–DG, CT–CA1, CT–CA3, DG–CA3, and CA1–CA3 showed significant increases for alpha-band coherences following 7-day CT-DBS. ^∗, ^∗∗, and ^∗∗∗ indicate significant differences in terms of coherence with p < 0.05, p < 0.01, and p < 0.001, respectively, relative to CT-DBS (baseline), analyzed by Wilcoxon signed-rank tests (mean ± SEM).

FIGURE 6

Comparsion of the two groups based on Western blotting results of the protein expressions of the DRD1, DRD2, and α4-nAChR receptors in the CA1, CA3, and DG. (A) The protein expressions of the DRD1, DRD2, and α4-nAChR receptors in the CA1, CA3, and DG in both sham controls and DBS group. (B) The mean normalized protein expression levels of the DRD1, DRD2, and α4-nAChR receptors in the CA1, CA3, and DG in the two groups. The CA1, CA3, and DG in the DBS group show significant increases in the protein expression levels of the DRD1 receptor but significant decreases in the levels of the DRD2 receptor compared with those in the sham controls. The CA1 and DG in the DBS group exhibited significant increases in the protein expression levels of the α4-nAChR receptor compared with those in the sham controls. ^∗, ^∗∗, and ^∗∗∗ indicated significant differences in terms of protein expression with p < 0.05, p < 0.01 and p < 0.001, respectively, relative to the sham controls, analyzed by Wilcoxon two-sample t-tests (mean ± SEM).