An automated platform for high-throughput mouse behavior and physiology with voluntary head-fixation

Abstract

Recording neural activity during animal behavior is a cornerstone of modern brain research. However, integration of cutting-edge technologies for neural circuit analysis with complex behavioral measurements poses a severe experimental bottleneck for researchers. Critical problems include a lack of standardization for psychometric and neurometric integration, and lack of tools that can generate large, sharable data sets for the research community in a time and cost effective way. Here, we introduce a novel mouse behavioral learning platform featuring voluntary head fixation and automated high-throughput data collection for integrating complex behavioral assays with virtually any physiological device. We provide experimental validation by demonstrating behavioral training of mice in visual discrimination and auditory detection tasks. To examine facile integration with physiology systems, we coupled the platform to a two-photon microscope for imaging of cortical networks at single-cell resolution. Our behavioral learning and recording platform is a prototype for the next generation of mouse cognitive studies.

Fig. 1

High-throughput automated setup for voluntary head fixation. a Head-restraining habituation system. Depiction of a mouse inside a transparent PVC tube drinking water from a water spout connected to a water tank (blue box). Narrowing rails on the sides of the tube progressively restrain the head-post (details in Supplementary Fig. 1). There is no latching mechanism for the head-post, hence mice can back out of the tube at any time. The containing box represents a mouse cage with an air-filtering lid and a metal grid for mouse containment (horizontal solid line under the lid) as in standard individually ventilated cages. Over the course of a couple of days, mice routinely self-restrained to drink water. b 3D view of the main dual cage setup (details of all hardware and software components in the Supplementary Material). c Side view of setup and labeling of its main components. d Side view of two setups (training capability of a single setup is four mice/day) housed in a mouse rack. Several setups can be accommodated in the same rack (current capacity: 12 platforms, 48 mice/day, ~12,000 trials/day). All units are in mm. Original technical drawings edited with permission from O’ Hara & Co., Ltd.

Fig. 2

Automated self-latching. Top-left panel shows a 3D rendering of the latching mechanism (details of the components and how they are assembled are provided in Supplementary Fig. 3). Panels 1–5 (circled numbers) show the sequence of steps leading to self-head fixation: (1) the head-plate (black bar on mouse head, Supplementary Fig. 1) is progressively restrained by narrowing rails (gray converging lines, Supplementary Fig. 3). (2) The forward motion of the head-plate mechanically lifts up the first pair of latching pins. (3, 4) The first pair of pins then lowers by gravity, and the continued forward motion of the animal similarly lifts up and down the second pair of latching pins, leading to the final self-head fixation (4). During 3, 4, small tilt and forward movements are allowed that reduce the probability of a “panic” response due to a sudden head fixation. (5) When the task session ends, a computer-controlled servo motor actuator lifts up both pairs of latching pins and releases the animal (Supplementary Fig. 4). Original technical drawings edited with permission from O’ Hara & Co., Ltd.

Fig. 3

Two-alternative forced choice visual discrimination task. a Example trial: a sinusoidal static grating with +45° clockwise (top) or −45° counterclockwise (cc, bottom) orientation from vertical was presented on the screen. Mice had to indicate whether the orientation was clockwise or cc by rotating with the front paws a wheel position between them and the screen, with the wheel rotation controlling the real-time close-looped orientation of the visual stimulus. b Trial structure of reward and negative feedback; ITI inter trial interval (randomized), OL open-loop period, CL close-loop period. c Changes in performance and number of trials over training days for an example mouse (m16058), colors reflect the left–right y-axes. Top horizontal bars and numbers indicate changes in the number of orientations the animal had to discriminate: two orientations (±45, a), up to six orientations for a minimum deviation from vertical of ±15°. The number of orientations increased when performance reached ~70%. d Same as c for a different example mouse (m16197). e, f Psychometric curves during three different learning phases, naive, initial learning, expert, for m16058 (e) and m16197 (f). Shaded area in the expert curves represent 95% bootstrapped confidence interval. g–i Changes in bias, lapsing rate, and slope derived from psychometric functions (and from a linear model for two orientation conditions, Methods) during learning (n = 8 mice). Each dot is an average across the first 6 days of the corresponding learning phase (Na, Le, and Ex for naive, learning, and expert). P-values are calculated from two-tailed Wilcoxon signed rank tests, **P < 0.01, *P < 0.05

Fig. 4

Go-no-go auditory detection task. a Example trial: in 70% of the trials, the animal was presented with brief tones (left panel). Mice had to respond by rotating the wheel in either direction. In the remaining 30% of trials, no tone was presented and the mouse had to refrain from rotating the wheel (right panel). b Trial structure of reward and negative feedback; ITI inter trial interval (randomized), OL open-loop period, CL close-loop period. c Box-and-whisker plot comparing percentage of Hit, Miss, false alarm (FA), and correct rejection (CR) responses of a representative mouse for first 10 sessions (Naive) and 15 consecutive sessions after the mouse reached a performance of d′ > 1.5 (learning). The whiskers show the minimum and maximum of data distribution; box lines show the 25th percentile, the median, and the 75th percentile. In naive mice, FA rates are higher than Hit rates possibly due to a startle reflex following the target sound presentation (P < 0.005, n = 10 sessions, Wilcoxon signed rank test)

Fig. 5

Compatibility of the setup with two-photon imaging. a Left panel: side view schematic of the latching unit for physiology with a ×10 objective lens for two-photon imaging. The two vertical gray segments indicate the latching pins (Supplementary Fig. 6). Right panel: 3D rendering of the latching part of the unit. b Wide-field GCaMP8 fluorescence image at the cortical surface. Red rectangle, ROI within V1. Scale bar 1 mm, dotted yellow lines and labels are segmented visual areas (Methods). c Repeated imaging of the same ROI across different days. Blue squares are magnified views of two example cells. Scale bar, 200 μm. d Visually evoked responses during the behavioral task recorded from cell1 shown in c. ΔF/F trial averages for left/right choices (red/blue lines). Shaded areas, 95% confidence interval. Time-out trials were not included in the data. e Responses to oriented gratings (30° diameter) from cell-2 shown in c. Error bars are s.e.m. Red curves, Gaussian fit to the data. Original technical drawings in a edited with permission from O’ Hara & Co., Ltd.