RIVETS: a mechanical system for in vivo and in vitro electrophysiology and imaging

Abstract

A number of recent studies have provided compelling demonstrations that both mice and rats can be trained to perform a variety of behavioral tasks while restrained by mechanical elements mounted to the skull. The independent development of this technique by a number of laboratories has led to diverse solutions. We found that these solutions often used expensive materials and impeded future development and modification in the absence of engineering support. In order to address these issues, here we report on the development of a flexible single hardware design for electrophysiology and imaging both in brain tissue in vitro. Our hardware facilitates the rapid conversion of a single preparation between physiology and imaging system and the conversion of a given system between preparations. In addition, our use of rapid prototyping machines ("3D printers") allows for the deployment of new designs within a day. Here, we present specifications for design and manufacturing as well as some data from our lab demonstrating the suitability of the design for physiology in behaving animals and imaging in vitro and in vivo.

Figure 1. RIVETS Head post and locking forks.

(a) RIVETS head post with removable target (b) RIVETS head post shown with narrow forks for implants in the x axes (c) RIVETS head post shown with wide forks for implants in the y axes.

Figure 2. Stability of RIVETS in awake, head-fixed mice.

(a) Two example images from a behavioral session with a mouse implanted with the RIVETs chamber. Video was collected at 100 Hz. (b) For each frame a line profile of intensity (inverted) was measured. Upper trace shows the mean (black) and standard deviation (gray) from line profiles collected for 3000 frames of a behavioral session. Example profiles spaced every 50 frames from early (red) to late (blue) profiles are shown. (c) From individual image profiles the correlation (upper), shift in the maximal location of cross correlation (middle), and the fractional difference between max correlation value and next greatest correlation score (lower) are shown for an example session with a single mouse. (d) Quantification of the profile correlations are shown for 5 mice in 5 sessions. Both histograms for inidivdual mice (light red) and the mean +/− standard deviation (red circles) are shown. (e) In all 5 mice no apparent shift was detected between frames as shown by the individual histograms (light blue) and means (blue circles).

Figure 3. In vivo electrophysiology using RIVETS.

(a) Schematic of the recording configuration used to obtain recordings shown in b–d. Briefly, a mouse trained to pull a lever (brass color) for water reward (blue water drop) was fixated in the RIVETS system under anesthesia. A small craniotomy was made and an 8 shank silicon probe array with 64 recording sites was lowered in to the dorsal striatum (depicted in the inset). Recordings were initiated ∼1 hour after recovery from anesthesia when the mouse was awake and actively moving the lever to obtain a water reward. (b) 2 minutes of a continuous recording from a single shank of an 8 shank silicon probe array is shown (‘electrodes’, light gray traces). Sorted spikes were colored according to the cluster to which they belonged and are indicated in overlay on the recording traces. Simultaneous monitoring of the lever position and beam breaks (‘licks’) are shown in the lower traces. Two types of events were detected and recorded in real time: threshold displacement of the lever (blue crosses) and delivery of water reward (red crosses). Rewards were delivered on a random subset (∼75%) of threshold crossings. (c) Averaged waveforms for the duration of the recording shown in b and separated by cluster (u01–u03) are shown for a window of −1 to +2 ms around the detected event. (d) The amplitude of each event in cluster u01 is projected into the space of electrodes 8 and 6 (el8 and el6) as a function of time (z-axis) for the entire 18 minutes of recording in this particular session. (e) Schematic of the recording configuration used to obtain recordings shown in f. As in the experiment depicted in a, a mouse was anesthetized and a small craniotomy was performed. The mouse was left lightly anesthetized while striatal neurons (∼2.5 mm below surface) were targeted from whole-cell patch clamping. (f) Representative whole-cell recording from a putative medium spiny neuron in the dorsal striatum. Left plot shows the response to a family of current steps from −160 to +160 pA in increments of 20 pA. Right plot shows 50 seconds of continuous recording showing a hyperpolarized resting potential and repetitive periods of sustained excitatory input.

Figure 4. Configurable stereotaxis, head post and slice chamber.

(a) RIVETS shown with simple one piece stereotaxis (b) RIVETS with customized two piece stereotaxis system (c) RIVETS with integrated stereotaxis insert for xy stage platform (d) RIVETS low profile optical head post (e) RIVETS in vitro slice chamber.

Figure 5. In vitro patch clamp electrophysiology and subcellular imaging using RIVETS.

To assess the stability of the slice chamber configuration we provide an example of 2 of the more challenging applications for in(a–b) First, we show that simultaneous whole-cell recordings from multiple nearby neurons can readily be obtained. In this case each neuron was filled with either a red (a; Alex594) or green (b; Alexa488) dye through the recording pipette. We further show that recorded neurons are sufficiently stable to perform detailed calcium-imaging experiments from the dendritic arbor. (c) A neuron was filled with both a red (Alexa594) dye and a calcium-sensitive fluorescent dye (Fluo-8). A complex linescan pattern covering multiple locations within the dendritic arbor was designed (left image, red overlay). The resulting calcium transient evoked by a barrage of backpropagating action potentials in shown in the colored traces at right. The colorization reflects progress in the linescan from beginning (cyan) to end (red). (d) To confirm the stability over many linescans the obtained structural image (red channel) was aligned for each line for a sequence of 250 lines. The maximum intensity trace and detected peaks for the image is shown in the middle plot. Red and blue crosses indicate the left and right edges of each detected structure, respectively. In the bottom plot we overlay the maximum intensity projection along the line for two sequences of line scans taken several minutes apart. The R² value of the correlation between these two maximum intensity projections was 0.99. Units of intensity are arbitrary.

Figure 6. In vivo two-photon imaging using RIVETS.

(a–c) A custom designed RIVETS holder was made to facilitate access of microscope objective (a). A craniotomy of ∼1.8 mm diameter was made and a small glass coverslip placed in the craniotomy (a–b). Imaging began by taking a low power (10×, 0.3 NA) of the surface of the brain (c). Upper image in (d) shows a volume reconstructed using a 60× (0.95 NA) objective to image layer V pyramidal neurons to ∼800 microns below the pial surface. Lower images show a single imaging plane collected 10 minutes apart. To assess stability of RIVETS small regions were imaged at 4–5 Hz. Single images taken 15 seconds apart (e) and the maximum intensity projection of the complete 300 frame video. To quantify stability both small ROIs (f–h) and whole frame correlations (i) were calculated. Z-axis displacement was estimated as changes in intensity of fluorescence of small ROIs. 4 example ROIs are shown in f over time. Histograms of all intensity values are plotted on right inset. Intensity histograms exhibited several forms ranging from highly consistent values through fluctuations in intensity (g). The distribution in peak heights of a multipeak histogram fits (h) reveal that most ROIs are stable across frames with a single dominant peak. Lateral shift can be estimated from the correlation between frames (i, upper plot) and the distance of the peak of the normalized cross-correlation (i, lower plot). Scale bars: a–b, 1 mm; c, 100 µm; e, 25 µm; f, 2.5 µm.