Simultaneous Cellular Imaging, Electrical Recording and Stimulation of Hippocampal Activity in Freely Behaving Mice

Abstract

Hippocampal sharp-wave ripple activity (SWRs) and the associated replay of neural activity patterns are well-known for their role in memory consolidation. This activity has been studied using electrophysiological approaches, as high temporal resolution is required to recognize SWRs in the neuronal signals. However, it has been difficult to analyze the individual contribution of neurons to task-specific SWRs, because it is hard to track neurons across a long time with electrophysiological recording. In this study, we recorded local field potential (LFP) signals in the hippocampal CA1 of freely behaving mice and simultaneously imaged calcium signals in contralateral CA1 to leverage the advantages of both electrophysiological and imaging approaches. We manufactured a custom-designed microdrive array and targeted tetrodes to the left hippocampus CA1 for LFP recording and applied electrical stimulation in the ventral hippocampal commissure (VHC) for closed-loop disruption of SWRs. Neuronal population imaging in the right hippocampal CA1 was performed using a miniature fluorescent microscope (Miniscope) and a genetically encoded calcium indicator. As SWRs show highly synchronized bilateral occurrence, calcium signals of SWR-participating neurons could be identified and tracked in spontaneous or SWR-disrupted conditions. Using this approach, we identified a subpopulation of CA1 neurons showing synchronous calcium elevation to SWRs. Our results showed that SWR-related calcium transients are more disrupted by electrical stimulation than non-SWRrelated calcium transients, validating the capability of the system to detect and disrupt SWRs. Our dual recording method can be used to uncover the dynamic participation of individual neurons in SWRs and replay over extended time windows.

Keywords: Brain wave; Calcium; Electrophysiology; Hippocampus.

Fig. 1

Design and structure of the microdrive. (A) Top view of a 3D CAD model of the microdrive with electrical interface board (EIB). a: Screw-driven shuttle for stimulation electrode targeting VHC. b: Screw-driven shuttle for tetrode targeting hippocampus CA1. c: Tetrode in the white matter for reference purposes. (B) Cross-section of one of the microdrive cannulas. The insertion depth of the stimulation electrode and tetrodes (yellow) were controlled by custom screws (black, arrow). (C) Close-up view of tetrodes emanating from the microdrive array and their target areas. (D) Cover of the drive for the protection of electrical components.

Fig. 2

Microdrive array fabrication materials and process. (A) Microdrive materials. Detailed list with providers is in the method section (2.3). (B~M) Microdrive array fabrication procedure. Detailed explanation is in the method section (2.3).

Fig. 3

Experimental timeline and recording system. (A) Experimental timeline. Single-photon calcium imaging: GRIN lens was implanted on the surface of the hippocampal CA1 of the right hemisphere (Day 1), and blood clearance at the surgical site was confirmed 2~3 weeks after the implantation (B). When calcium activity was observed through the implanted lens, mice were equipped with a baseplate (Day 27) to mount the miniscope (D). Once microdrive implantation was consolidated, miniscope was mounted on the baseplate and data acquisition was performed simultaneously with the LFP recording (Day 28~50) (E). Electrophysiological recording: Microdrive array manufacturing takes 1 day (8 hours, Day 13). Assembled microdrive array was carefully implanted above the target brain regions (Day 14) (C). After implantation, animals underwent 5~7 days of post-operative recovery. After recovery, electrophysiological signals were acquired, together with the calcium signal (Day 28~50) (E). In the home cage animals only carried the microdrive array, but not the miniscope, which was mounted onto the baseplate just prior to the start of the recording session. After experiments were completed, the implantation sites for the GRIN lens and tetrodes were validated histologically (Day 50~60) (F). (G) Set up of the acquisition system for simultaneous imaging, electrical recording, and closed-loop stimulation. 1-photon calcium signal was acquired by miniscope connected to a dedicated computer using Miniscope Controller software (a). Electrical signals were acquired with a Digilynx acquisition system and stored on a computer running Cheetah software (b). The digitized electrical signals were also streamed to a separate computer running Falcon software for real-time hippocampal ripple detection (c) and closed-loop feedback stimulation in the VHC (d). Each video frame from the miniscope was timestamped in the Digilynx system for synchronization with the electrical signals (e). (H) Recording chamber. (I) Coronal section of the dorsal hippocampus showing the location of the implanted GRIN lens. (J) Coronal section of the dorsal hippocampus showing a tetrode track. The arrow indicates the lesion made by an electrical current applied to the tetrode to verify the recording location. (K) Accumulated time deviation between acquired frames and theoretically generated frames of calcium imaging (Black: acquired frames, Red: timepoint of dropped frames, Blue: Corrected frames).

Fig. 4

Electrophysiological signals and calcium transient were simultaneously acquired in spontaneous and SWR-disrupted conditions. All data from one representative animal (Animal 1). SP: average peak amplitude in spontaneous condition, ST: average peak amplitude in SWR-disrupted condition. (A) Representative wide-band filtered (1~6000 Hz, top) and ripple-band filtered (140~225 Hz, bottom) hippocampal LFP signal in the spontaneous condition (left) and SWR-disrupted condition (right). (B) Average spontaneous SWR power (left) and disrupted SWR power (right, dashed) (red: SWR detection). (C) Representative calcium signals in the spontaneous condition (left) and SWR-disrupted condition (right) (red: SWR detection, blue: detected peaks, gray: baseline).

Fig. 5

Different subsets of cells respond to SWRs depending on the behavioral state. Calcium peaks from all cells were classified into two categories depending on the presence of preceding SWRs. (A, C): SWR-preceding peaks, (B, D): non SWR-preceding peaks. (A) Proportion of units by the SWR-preceding peak property from all units. Left: proportion of units where the peak amplitude decreased in the SWR-disrupted condition. Right: proportion of units where the peak amplitude increased in the SWR-disrupted condition. (Total units: 320, p-value=4.26e-24, solid: bootstrapped 99% confidence interval). (B) Proportion of units by the non SWR-preceding peak property from all units. Left: proportion of cells for which the peak amplitude decreased in the SWR-disrupted condition. Right: proportion of units where the peak amplitude increased in the SWR-disrupted condition. Proportions of Fig. 5A and Fig. 5C were statistically compared using a two proportion z-test (p-value<0.001). (Total units: 320, p-value=2.08e-05, solid: bootstrapped 99% confidence interval). (C) Proportion of units by the SWR-preceding peak property from individual animals (Animal 1: 105 units, animal 2: 122 units, animal 3: 47 units, animal 4: 46 units). (D) Proportion of units by the non SWR-preceding peak property from individual animals (Animal 1: 105 units, animal 2: 122 units, animal 3: 47 units, animal 4: 46 units). (E) Distribution of average calcium peak amplitude ratio of SWR-disrupted condition compared to the spontaneous condition of peaks with preceding SWRs (red) and peaks without preceding SWRs (gray) of 90% of units during the whole recording (Wilcoxon rank-sum test, p<0.0001, dashed: median, solid: bootstrapped 99% confidence interval. The gray line indicates where the average calcium peak amplitude in the spontaneous condition is equal to the average calcium peak amplitude in the SWR-disrupted condition).