Asymmetric visual input directs larval zebrafish motor bias (A) Diagram of the CP assay. The inset shows an overhead view of fish and mounting orientations (RD, right side down; UP, both eyes up; LD, left side down). (B) Timeline of the CP assay for inducing turn bias direction. A green bar indicates the duration of embedding and visual control. (C) Representative tracks of mounted individuals during baseline (left) and following loss of illumination (dark, right). Scale bars, 10 mm. The bar on the right shows time in seconds. (D–G) Match index for mounted individuals during baseline (yellow) and dark (gray) recording conditions. A dotted red line indicates the random match index. A circled asterisk indicates p < 0.05, one-sample Wilcoxon signed rank test to 0.5 (RD, N = 40; LD, N = 35; UP, N = 62). Also shown are relative frequencies of paired directional transitions from the first two trials compared with the last two for (E) upright, (F) left-mounted, or (G) right-mounted fish. Numbers for the first two trials are denoted in colored circles for individuals that had repeated left (cyan), right (magenta), or random (orange) trials. Paired arrows in circles indicate turn direction over the first two trials. Arrows between circles indicate transition frequency, and adjacent numbers are percentages. (H–K) Average BR (averaged over four trials per individual) for individuals in the CP. A box above each graph shows dpf (days), and the green bar indicates the time in the CP assay. Bottom: boxes show the ratio of right-biased (+BR) to left-biased (−BR) individuals per group. All CP assay BR graphs show these features. (H) CP assay from 2–4 dpf (LD, 39; RD, 40; UP, 62) with light overhead. (I and J) Same as (H) with the light source rotated below the embedded larvae (I) (LD, N = 51; RD, N = 53; UP, N = 42) or with no photic experience (J) (LD, N = 44; RD, N = 28; UP, N = 54). (K) CP assay performed later in development compared with (H)–(J). Behavior was tested at 8 dpf (LD, N, 31; RD, N = 28; UP, N = 39). ∗p < 0.05, one-tailed t test to 0. White circles with a black vertical line show the group mean and SEM. (L) Diagram of visually mediated changes to turn bias direction during the CP. All error bars represent ±SEM.

y279-specified AMNs are GABAergic Fluorescence in situ hybridization (FISH) HCR of Tg(y279:Gal4; UAS:kaede) larvae at 3 dpf. All images are single-plane confocal scans showing AMNs from a single hemisphere. (A) Kaede mRNA (left, green), GAD1b mRNA (center, magenta), and merged (right). A dotted circle outlines a single cell body, and arrows indicate co-localization. Similar mRNA labeling was observed across fish: GAD2 (observed in 3 larvae) (B) and VGlut2b (observed in 4 larvae). (C) A similar pattern was observed in 3 larvae for each label. Scale bar, 10 μm.

Thalamic neurons show functional asymmetries correlated to motor bias (A–C) Outline of steps for calcium imaging experiments. (A) Larvae were prepped in the CP assay prior to calcium imaging. (B) Experiment outline showing the time period of the CP assay (mount, green), behavior testing to confirm turn bias (test, blue), and calcium imaging (Ca2+, red). (C) Schematic of calcium imaging set up on 2P and photo-stimulation series. (D) Representative segmentation of lateral thalamic (yellow), greater thalamic (purple), PT (pink), and pallium (green) neurons using Tg(elavl3: H2B-GCaMP6f; y279 Gal4; UAS: nsfB-mCherry). Scale bar, 20μm. (E) Example of matched and opposed hemispheres for a right-biased individual. (F) Raster plot showing responses from a subset of lateral thalamus neurons. Vertical gray bars indicate light transition (0–60 s light ON, 61–120 s light OFF, 121–180 s light ON). Color scale, fluorescence change (ΔF/F). (G) Proportions of matched and opposed visually evoked ON and OFF responses per region. Chart sizes are scaled to represent the proportion of neurons represented. A cross indicates groups with 50 or fewer neurons, which are set to a standard minimal scale. (H and I) Spatial distribution for matched and opposed responses between hemispheres (OFF response, orange, N = 699; ON response, purple, N = 216). A gray outline indicates the position of neurons with no light response. OFF response mean and standard error for (I) lateral thalamic y279+ (matched, N = 75; opposed, N = 104), (J) lateral thalamic y279− (matched, N = 28; opposed, N = 38), and (K) greater thalamic (matched, N = 61; opposed, N = 79), and (L) PT regions (matched, N = 167; opposed N = 143). An asterisk with a bar indicates a two-tailed t test p < 0.05 for at least 10 consecutive time points. (M) Diagram of unilateral 2P ablation. (N) Average BR of larvae with left (blue) or right (orange) hemispheric ablation of AMNs (left, N = 14; right, N = 17), medial thalamic neurons (left, N = 15; right, N = 14), and unablated controls (gray, N = 36). (O) Absolute BR of control (N = 36), AMN (N = 31), and medial thalamic neuron (N = 29) ablation. ∗p < 0.05 to 0. An asterisk with brackets indicates p < 0.05 between groups. All error bars and envelopes on line graphs represent ±SEM.

AMNs respond to visual input (A) Schematic of unilateral enucleation. (B) Representative image showing a single confocal plane for thalamic neuron imaging. (C) Illustrative example showing regions for analysis in (D)–(F) (lateral thalamus [L. Th; white], greater thalamus [Gr. Th; green], and AMN neuropil [Npil; orange]). Scale bar, 20 μm. Shown are standardized calcium responses to visual stimuli in unilaterally enucleated larvae. Vertical dotted lines denote the time of light OFF (60 s) and light restoration (120 s). Solid lines are an average, and envelopes show ±SEM. (D) Calcium responses in the L. Th (N = 11) in the intact (red) and enucleated (purple) hemisphere. (E) Same as (C), showing Gr. Th (N = 11) responses. (F) iGluSnfer responses from intact (red) and enucleated (purple) hemispheres recorded from the adjacent AMN Npil (N = 9). (G) Representative image showing a single confocal plane for optic tectum imaging. (H) Illustrative region used for recording calcium responses in the optic tectum Npil. Scale bar, 50 μm. (I) Npil responses for hemispheres matched to the direction of turn bias (red) or opposed (purple) (N = 13). (J) Npil responses contralateral to the enucleated eye (red) or ipsilateral (purple) (N = 11). Asterisk and bar, two-tailed t test, p < 0.05 for at least three consecutive time points. Envelopes on line graphs represent ±SEM.

GABAergic signaling controls CP timing (A) Average BR of larvae following asymmetric visual experience. A schematic (top box) shows the time in dpf (days) of CP assay mounting (green) and drug treatments (orange). Control (LD, N = 32; RD, N = 35), PTZ (LD, N = 30; RD, N = 31), and muscimol + PTZ (LD, N = 26; RD, N = 27). Bottom R/L values show the proportion of left BR (negative values) to right BR (positive value) individuals. A red line indicates random movement. (B) Same as in (A) with different activity-modulating drug treatments (control: LD, N = 46; RD, N = 47; 500 μM caffeine: LD, N = 37; RD, N = 43; 400 μM 4-AP: LD, N = 30; RD, N = 29). (C) Representative c-fos mRNA labeling following drug treatments. Shown are single-plane confocal scans of 4-dpf larvae. An arrow indicates areas of increased c-fos expression. A yellow outline shows the area analyzed for c-fos signal quantification. Scale bar, 100 μM. (D) Quantification of (C). Shown is the average mean intensity of c-fos from untreated controls (N = 41), PTZ (N = 16), muscimol + PTZ (N = 14), caffeine (N = 14), and 4-AP (N = 12). Values are standardized to the control. (E) Same as in (A) with larvae prepped in the CP assay from 4–6 dpf (see timeline in the box above) (control: LD, N = 39; RD, N = 33; PTZ: LD, N = 33; RD, N = 30). ∗p < 0.05 to 0.0. An asterisk with lines in (D) shows p < 0.05 between groups. (F–H) Frequency of photo-mediated responses for all regions in PTZ or control individuals (tan, no response [NR]; orange, OFF response; red, ON response; purple, OFF/ON response). Shown is the average OFF response for the matched (purple) or opposed (red) hemisphere based on positioning in the CP assay and expected turn bias direction for (G) control (matched, N = 123 neurons; opposed, N = 43 neurons from N = 19 larvae; N = 12 RD and N = 7 LD) and (H) PTZ-treated larvae (matched, N = 127 neurons; opposed, N = 110 neurons from N = 19 larvae; N = 12 RD and N = 7 LD). An asterisk with a bar indicates a two-tailed t test p < 0.05 for at least 10 consecutive time points comparing matched and opposed responses. All error bars and envelopes on line graphs represent ±SEM.

AMN activation drives motor responses (A) Schematic of the 2P setup for optogenetic stimulation and tail recording (left) and a diagram of head-embedded larvae and field of view containing AMNs (right). (B) Photic stimulation series and wavelengths. (C) Average tail displacement (mm) of UAS:TagRFP-v2a-paGFP (control) larvae following no-stimulation dark controls (gray), 1,040-nm (green), and 940-nm laser (blue) exposure. Individuals tested, N = 13. (D) Representative tail traces for analysis in (C) during dark control, 1,040-nm, and 940-nm recording. (E and F) Optogenetic inhibition of AMNs using GtACR2, showing tail displacement (E) and illustrative tail trace examples (F) (N = 8). (G and H) Same as above for UAS:ChEF-v2a-mCherry-injected individuals (N = 10). An asterisk with lines shows p < 0.05 between groups. All error bars represent ±SEM.

Circuit diagram Shown is a representative diagram of the proposed circuit model and visual experience impact on turn bias. Excitatory connections are shown with circled ends and inhibitory connections with flat ends. Examples of visual plasticity are shown for right-eye, both eyes (center), and left-eye stimulation conditions. The stimulated eye is shown in yellow. Line thickness indicates response strength. We propose that asymmetric descending AMN disinhibitory signaling to premotor neurons releases motor activity to establish turn bias.