Fish-on-Chips 2.0 for studying chemosensation in larval zebrafish.

(a) Upper panel: Schematic of the valence assay. Zebrafish larvae swimming in a two-dimensional arena (40 mm × 40 mm × 1.5 mm) are imaged at 20 fps under infrared (IR) illumination in the absence of visible light (see Supplementary Figure 1a for a more detailed schematic for the setup). Four quadrants with equal area are created and maintained by a constant, slow inflow of fluid at each corner (marked by arrows), with outflow occurring at a shared central outlet. The laminar flow maintains static borders between the zones. Lower panel: Visualization of the well-defined quadrants by infusing IR dyes into quadrant II and IV, and water into quadrant I and III. Scale bar: 5 mm. (b) Assayed chemicals and tested concentrations (varied across 3 consecutive orders of magnitude for each), including valine (Val), alanine (Ala), prostaglandin F2α (PGF2α), cadaverine (Cad), putrescine (Put), chenodeoxycholic acid (CDCA), glycodeoxycholic acid (GDCA), adenosine (Ade), and sodium chloride (NaCl). These were dissolved either in water or 0.5% DMSO and infused into quadrant II (lowest concentration) to IV (highest concentration). (c) Differences in % time spent in the chemical quadrants (i.e., II– IV, with different and increasing concentrations for each of the tested chemicals as indicated in (b) compared with the control quadrant (i.e., I), showing the values of individual assays (gray dots) and the medians (black-outlined dots). Below each chemical name, the first number indicates the total number of larvae assayed, and the second number specifies the total number of assays performed. P-values: Multiple Z-tests comparing different quadrant combinations’ means to zero, with population standard deviation (S.D.) estimated from the sample S.D. of the control group (see Methods). Statistical significance after Benjamini-Hochberg adjustment: Black, significant; gray, near-significant. (d) Left panel: A drawing of the enhanced polydimethylsiloxane (PDMS)-based microfluidic device, featuring multiple chemical delivery channels on both sides of the larval head chamber. These channels are offset from the front to accommodate front excitation light sheet scanning, in addition to the right side light sheet. Right panel: A close-up illustration of the neural-behavioral imaging capabilities in a larval subject during precisely controlled chemical stimulus presentation. See Supplementary Figure 2 for more details. (e) Pre-event intervals (PEIs) [1/ frequency] of spontaneous saccade and tail flip for all recorded events in the microfluidic device. Horizontal lines show the median, 25th and 75th percentiles for each group. n = 5 larvae with 216 spontaneous saccades and 1018 spontaneous tail flips, each with a recorded preceding reference event. Shaded areas scale according to the probability density function of values. P-value: Two-sided Wilcoxon rank-sum test. Dotted rose lines indicate the respective reduced frequencies in typical studies utilizing agarose-based embedding, both being < 0.1 Hz (see refs 50,74).

Tail flip dynamically aligns with saccade in rate and direction for performing spontaneous body turns.

(a) Schematic illustrating the various temporal parameters of an example of coupled saccade-tail flip (S-T) event. (b) & (c) Histograms showing (b) onset delay and (c) offset delay of coupled saccade-tail flip (S-T) events. Note that the precision of measurements for the onset and offset for saccade and tail flip were 0.0625s and 0.0025s due to the respective sampling rates. Black and gray dashed lines indicate median and zero respectively. P-values: Two-sided Wilcoxon signed-rank test. (d) & (e) Pre-event intervals [1/ frequency] of self and flanking (d) saccades and (e) tail flips with respect to independent (S, saccade; T, tail flip) or S-T events. (f) Pie chart showing the proportions of contra- and ipsilateral turns, comparing the directionality of saccades (S) and tail flips (T; only turns considered, see Methods) in S-T events. P-value: Two-sided binomial test. (g) & (h) Laterality indices (defined as proportions of ipsilateral turns) in flanking (g) saccades and (h) turn tail flips (tail flip asymmetry ≥ 1.5°) with respect to independent (S, T) or S-T events. In (h), the raw proportions (thin lines) and corresponding moving-average curves (thick lines) are shown. P-value: One-sided Chi-squared test comparing peaks and ± 1 events. (i) Tail flip asymmetry of self and flanking tail flips for independent tail flips (T) or S-T events with tail flip asymmetry ≥ 1.5° (see Supplementary Figure 3k for all events). (j) Average undulation of self and flanking tail flips for independent tail flips (T) or S-T events with tail flip asymmetry ≥ 1.5° (see Supplementary Figure 3l for all events). (k) Illustration of the temporal, directional and kinematic coordination of spontaneous saccades and tail flips. The temporal and directional alignment of tail flips with saccades, as well as their turning intent gradually decreases following a single S-T event and gradually increases preceding a single S-T event. Abbreviations in (b)–(c), ON, onset; OFF, offset. In (d), (e), (i) & (j), the thin curves represent the medians, 75th and 25th percentiles of all events. The raw median values (thin lines) and corresponding moving-average curves (thick lines) are shown in all these plots except (d). P-values: Two-sided Wilcoxon rank-sum test comparing peaks and ± 1 events. For all plots, data were collected from n = 5 larvae, with 330 spontaneous saccades and 1,116 spontaneous tail flips.

Freely moving zebrafish larvae dynamically align bouts in rate and direction for spontaneous body turns.

(a) Medians and the 95% confidence intervals (based on bootstrapping with 1000 resamples) of the turn angle of all swim bouts vs. their pre-bout intervals (binned). Gray dots are individual data points. Turn angle values (sampled at an angular resolution of 10°) are slightly vertically dispersed for visualization. Spearman’s rank correlation coefficients (r) and associated P-values are shown. (b) Pre-event interval [1/ frequency] of self and flanking swim bouts for non-turn (teal) or turn (coral) swim bouts (see Methods). Thin curves represent the raw median values, 75th and 25th percentiles. Thick curves are the moving-average of medians. (c) Pie chart showing the proportions of contra- and ipsilateral consecutive turns (i.e., comparing each swim bout to its preceding event). P-value: Two-sided binomial test. (d) Laterality index (defined as proportion of ipsilateral turn) in flanking swim bouts for non-turn (teal) or turn (coral) events. The raw median values (thin lines) and corresponding moving-average curves (thick lines) are shown. P-values: One-sided Chi-squared test comparing peaks and ± 1 events. (e) Turn angle of self and flanking tail flips for non-turn (teal) or turn (coral) swim bouts. The thin curves represent mean ± S.E.M. across time. The raw (thin lines) and corresponding moving-average curves (thick lines) for the mean are shown. In (b) & (e), P-values: Two-sided Wilcoxon rank-sum test comparing peaks and ± 1 events. Note that however in (e), the centered bout (i.e., 0th next bout) is by definition different and therefore only flanking bouts are compared. For all plots, data were collected from n = 11 larvae, with 1,529 spontaneous swim bouts.

Aversive but not appetitive chemosensory cues amplify eye-body coordination via modulations of individual and joint movements of saccades and tail flips.

(a) Illustration of chemical stimulation paradigm. During each trial, one appetitive or aversive chemical was administered to a tethered larval subject under neural-behavioral imaging. (b) Histograms showing the distributions of difference in saccade count between the 10-second stimulus window and the 10-second pre-stimulus period, across trials for each chemical and blank control. Each dot represents a trial-averaged saccade count change from a larva (dispersed vertically for visualization). (c) & (d) Stimulus-evoked changes in (c) saccade frequency and (d) normalized tail flip frequency (see Methods), compared with the pre-stimulus average. (e) Pre-tail flip interval [1/ frequency] of chemical stimuli-associated tail flips. (f) Stimulus-associated changes in the relative frequency of coupled saccade-tail flip (S-T) events vs. independent tail flips (i.e., Δ[S-T event frequency - independent tail flip frequency]) across time. (g) & (h) Kinematic parameters of chemical stimuli-associated tail flips including (g) tail flip asymmetry (all tail flips) and (h) average undulation (for tail flips with ≥ 4 undulation cycles). (i)–(m) Scatter plots of tail flip asymmetry vs. various temporal parameters for S-T events, including saccade-tail flip (i) onset delay, (j) offset delay, (k) overlapping time interval, (l) saccade duration, and (m) tail flip duration. Individual data points are shown in squares, diamonds and circles for the blank, appetitive and aversive groups respectively. (n) & (o) Illustrations of larval zebrafish behavioral changes upon (n) aversive and (o) appetitive chemicals presentations. In (c), (d), and (f), quantities are binned and normalized to the average values in the 10-second pre-stimulus windows (see Methods). Lines and shadows show mean ± SEM (across trials). In (d), the raw means (thin lines) and corresponding moving-average curves (thick lines) are shown. Dashed rectangles mark the stimulus window. In (e), (g), and (h), horizontal lines indicate medians, 75th and 25th percentiles. Shadows of the violin plots scale according to the probability density function. In (b)–(h), P-values: Kruskal–Wallis test with Dunn-Sidak post hoc test. Eta squared (η2) is provided as a measure of effect size. In (i)–(m), lines show best linear fits of each data group and shadows show 95% prediction confidence intervals of the aversive chemical data group (the only that exhibited significant associations). Pearson’s correlation coefficients (r) and associated P-values are shown for the aversive chemical data points. Only data points with saccade onset certainly earlier than the tail flip onset were included in the regression analysis (see Methods). In (l), saccade duration values (sampled at a temporal resolution of 0.125s) are slightly horizontally dispersed for visualization. Abbreviations: S, saccade; T, tail flip; ON, onset; OFF, offset; Ap, appetitive; Av, aversive. For all plots, data were collected from n = 5 larvae, and the total numbers of blank control, appetitive and aversive chemical trials used in statistical analyses are 28, 52 and 53, respectively. In (d), the numbers of blank, appetitive and aversive trials with non-zero pre-stimulus tail flip frequency used in statistical analyses are 19, 34 and 27, respectively. In (e), the total numbers of blank, appetitive and aversive events used in statistical analyses are 69, 130 and 89, respectively. In (g), the total numbers of blank, appetitive and aversive events used in statistical analyses are 89, 167 and 124, respectively. In (h), the total numbers of blank, appetitive and aversive events used in statistical analyses are 76, 156 and 111, respectively.

Pallial neuronal subsets are highly co-tuned to coupled saccade-tail flip (S-T) events and chemical aversiveness.

(a) Heatmap of brain-wide spontaneous neuronal activities of a larval subject, acquired simultaneously with the motor outputs shown in (b). (b) Spontaneous saccades and tail flips of the same larval subject in (a), with a zoom-in example tail flip shown below. Temporally coupled saccade-tail flips (≤ 0.5s onset time difference) are indicated with coral blocks. Other saccades and tail flips are highlighted with blue and teal blocks, respectively. (c) Mean intensity projections (to transverse and sagittal planes) of motor preference of motor-encoding regions-of-interest (ROIs) (see Methods, x-axis in (g)). (d) Regional mean ± S.E.M. (across larvae) of motor preference. (e) Mean intensity projects (to transverse and sagittal planes) of valence preference for all valence-encoding ROIs (see Methods, y-axis in (g)). (f) Regional mean ± S.E.M. (across larvae) of valence preference. (g) Scatter plot of valence preference vs. motor preference of the union set of motor- and valence-encoding neurons. The four quadrants classify neurons into different sensorimotor tuning properties. (h) Mean intensity projections (to transverse and sagittal planes) of norm of sensorimotor preference (see Methods), for all ROIs preferring S-T event and aversive valence (i.e., quadrant I in (g), also see Supplementary Figure 7c for zoom-in of the forebrain area). (i) Regional mean ± S.E.M. (across larvae) of norm of sensorimotor preference for all ROIs preferring S-T event and aversive valence. (j) Mean intensity projections (to transverse and sagittal planes) of norm of sensorimotor preference for all ROIs preferring independent tail flip and appetitive valence (quadrant III in (g), also see Supplementary Figure 7d for zoom-in of the forebrain area). (k) Regional mean ± S.E.M. (across larvae) of norm of sensorimotor preference, for all ROIs preferring independent tail flip and appetitive valence. Scale bars in (c): 50 μm in the Z-Brain Atlas space. In (d), (f), and (i), P-values: Right-sided t-test comparing the pallium and non-pallial brain regions. In (k), P-value: Left-sided t-test comparing the pallium and non-pallial brain regions. In (c)–(f) & (h)–(k), the major brain regions (Tel: telencephalon, Di: diencephalon, Me: mesencephalon and Rh: rhombencephalon) are color-coded. Abbreviations of brain regions: OB olfactory bulb, Pa pallium, sPa subpallium, PO preoptic area, PT posterior tuberculum, dTh dorsal thalamus, vTh ventral thalamus, rHT rostral hypothalamus, iHT intermediate hypothalamus, pTec pretectum, Teg tegmentum, Rh1 rhombomere 1, Rh2 rhombomere 2, Rh3 rhombomere 3, Rh4 rhombomere 4, Rh5 rhombomere 5, Rh6 rhombomere 6, Rh7 rhombomere 7, cHb caudal hindbrain. Other abbreviations: S-T, coupled saccade-tail flip; T, independent tail flip; Ap, appetitive; Av, aversive. For all plots, the neural activity data correspond to the behavioral data reported in Figures 2 & 4, collected from n = 5 larvae.

Pallial motor neuron activity onset precedes other brain regions during chemical cue-associated coupled saccade-tail flip (S-T) but not independent tail flip (T) events.

(a) Upper panel: Mean ± S.E.M. (across larvae) responses of motor-encoding ROIs in the pallium (left) and non-pallial brain regions (right) to chemical cue-associated S-T events. Lower panel: Mean ± S.E.M. (across larvae) responses of motor-encoding ROIs in the pallium to chemosensory cue-associated S-T events with saccade-tail flip onset delay < 0.1 seconds (left) or ≥ 0.1 seconds (right). (b) Mean ± S.E.M. (across larvae) responses of motor-encoding ROIs in the pallium (left) and other brain regions (right) to chemical cue-associated independent tail flips. In (a) & (b), dotted traces show the corresponding neuronal responses of spontaneous events. Dotted vertical lines indicate the time points of tail flip onsets (in a corresponding neuronal imaging frame). The neural activity data correspond to the behavioral data reported in Figures 2 & 4, collected from n = 5 larvae. P-values: Right-sided Z-tests comparing pre-chemical cue-associated event neuronal activity at each time point to zero. Significance was set at 2.5 standard deviations (S.D.), with S.D. estimated from spontaneous event baseline activity variability. Adjustment for multiple comparisons across time points and brain regions was made using the Benjamini-Hochberg procedure (see Methods). (c) Schematic illustrations showing the unique role of pallium in representing the transformation of negative chemical valence into initiation of amplified body-turning movements. Red and blue arrows represent the neuronal information flow for the processing of negative and positive chemical cue valence, respectively. Abbreviation: OSN, olfactory sensory neurons.

List of motor regressors and mutual information measures.

Schematics and additional analysis of the valence determination assay.

(a) Schematic of the valence determination assay. Zebrafish larvae swimming in a two-dimensional arena (40 mm × 40 mm × 1.5 mm) are imaged at 20 fps under infrared (IR) illumination in the absence of visible light. Four quadrants with equal area are created and maintained by a constant, slow inflow of fluid at each corner, with outflow at a shared central outlet. The laminar flow maintains static borders between the zones. (b) & (c) Differences in % time spent in the chemical quadrants (i.e., II–IV, with different and increasing concentrations for each of the tested chemicals as indicated in Figure 1b) compared with the water quadrant (i.e., I), showing the values of individual assays (gray dots) and the medians (black-outlined dots). The experiments were performed with either (b) 0.5% DMSO or (c) water in quadrant I. Below each chemical name, the first number indicates the total number of larvae assayed, and the second number specifies the total number of assays performed. P-values: multiple Z-tests across different quadrant combinations’ means to zero, with population standard deviation (S.D.) estimated from the sample S.D. of the control group (see Methods). Statistical significance after Benjamini-Hochberg adjustment: Black, significant; gray, near-significant.

More details on Fish-on-Chip 2.0: Enhanced platform for naturalistic behavioral-neural readouts in tethered larval zebrafish.

(a) A photograph of a tethered larval zebrafish loaded into a PDMS-based microfluidic device (scale bar: 1 mm). Its eyes and tail are free to move while the head is fixed in place for fluorescence imaging. (b) Median pre-event intervals (PEIs) [1/ frequency] of spontaneous saccade and tail flip of each larva. Horizontal lines show the median, 25th and 75th percentiles for each group. n = 5 larvae with 216 spontaneous saccades and 1,018 spontaneous tail flips, each with a recorded preceding reference event. (c) Optical layout of the light sheet microscope, including two light sheet scanning modules which scan the larval brain from the side and the front respectively, a fluorescence detection module, and an IR imaging module for saccade and tail movements recording. (d) Upper panel: A fluorescence image showing sodium fluorescein in the microfluidic device excited by the front and side scanning light sheets (LS, illustrated by overlaid blue shadows), with photomasks placed to avoid the light sheets from directly reaching the larval eye positions (marked with white dashed ovals). A 250-µm window in the front photomask allows for the passage of the front excitation light to scan the brain regions between the eyes. Scale bar: 100 µm. Lower panel: A photograph showing the microfluidic device, objectives, and the associated tubings. (e) Left panel: An example functional image plane showing neural activation at two time points: before and after the presentation of 1 mM cadaverine. Anatomical reference images are overlaid. Right panel: A maximum intensity projection (MIP) and the time-series images of the first 30-second interval, focusing on a zoomed-in forebrain region indicated (dashed white rectangle on the left). 6 regions-of-interest (ROIs) are outlined. The red arrow marks cadaverine stimulus onset. White arrows indicate the locations and times of each ROIs near a calcium event’s peak. Scale bars: 100 μm. (f) The corresponding calcium signal traces of 38 representative neurons recorded before and after the presentation of 1 mM cadaverine, with onset indicated by the red arrow and dotted red line. Scale bar: 10 seconds. (g) Summary of key differences between Fish-on-Chip 1.0 and 2.0. The enhanced 2.0 platform eliminates the need for tilting or right-eye ablation in larvae, allowing for naturalistic, whole-body and brain-wide behavioral-neural recordings during spontaneous behavior and chemosensory cue presentations.

Additional analysis of spontaneous saccade-tail flip coordination.

(a) & (b) Histograms showing (a) overlapping time interval and (b) difference in onset-offset delay of coupled saccade-tail flip (S-T) events. Note that the precision of measurements for the onset and offset for saccade and tail flip were 0.0625s and 0.0025s due to the respective sampling rates. Black and gray dashed lines indicate median and zero respectively. P-values: Two-sided Wilcoxon signed-rank test. (c) Venn diagram showing all counts of spontaneous saccades, tail flips and S-T events (n = 5 larvae). (d) & (e) Pie charts showing the proportions of contra- and ipsilateral turns, comparing the directionality of (d) saccades and their preceding events, and (e) tail flips and their preceding events. P-values: Two-sided binomial test. (f) Illustration of the various kinematic parameters of an example tail flip. Dashed lines indicate the right peak mean angle, start-end point mean angle, and left peak mean angle. (g)–(j) Medians and 95% confidence intervals (based on bootstrapping with 1000 resamples) of all tail flips’ (g) tail flip asymmetry vs. their pre-tail flip interval [1/ frequency] (binned), (h) tail flip asymmetry vs. their magnitude (binned), (i) magnitude vs. their pre-tail flip interval [1/ frequency] (binned) and (j) average undulation vs. their pre-tail flip interval [1/ frequency]. Gray dots are individual data points. Spearman’s rank correlation coefficients (r) and associated P-values are shown. (k) Tail flip asymmetry of self and flanking tail flips for all S-T events or all independent tail flips (T) regardless of tail flip asymmetry. (l) Average undulation of self and flanking tail flips for all S-T or all T events regardless of tail flip asymmetry. In (k) & (l), the thin curves represent the medians, 75th and 25th percentiles of all events. P-values: Two-sided Wilcoxon rank-sum test comparing peaks and ± 1 events.

Temporal and directional characteristics of coupled saccade-tail flip (S-T) events under chemical stimulus presentation.

(a)–(d) Histograms of saccade-tail flip (a) onset delay, (b) overlapping time interval, (c) offset delay and (d) difference in onset-offset delay of S-T events during blank stimuli. Note that the precision of measurements for the onset and offset for saccade and tail flip were 0.0625s and 0.0025s due to the respective sampling rates. Black and gray dashed lines indicate median and zero respectively. P-values: Two-sided Wilcoxon signed-rank test. (e)–(g) Pie charts showing the proportions of contra- and ipsilateral turns, comparing the directionality of (e) saccades and turns in S-T events, (f) saccades and their preceding events, and (g) tail flips and their preceding events. P-values: Two-sided binomial test. (h)–(n) and (o)–(u) are similar to (a)–(g) but showing results during appetitive and aversive chemical presentations, respectively. Abbreviations: S, saccade; T, tail flip; ON, onset; OFF, offset.

Additional kinematic and temporal analysis of spontaneous coupled saccade-tail flip (S-T) coordination during chemical presentation.

(a)–(c) Kinematic parameters of tail flips subgrouped into chemical stimulus-associated pre-S-T (independent), S-T and post-S-T (independent) tail flips including (a) tail flip asymmetry, (b) average undulation, and (c) pre-tail flip interval [1/ frequency]. Horizontal lines indicate the medians, 75th and 25th percentiles. Shadows of the violin plots scale according to the probability density function. P-values: Kruskal–Wallis test with Dunn-Sidak post hoc test. Eta squared (η2) is provided as a measure of effect size. Abbreviations: Ap, appetitive; Av, aversive. (d)–(h) Scatter plots of average undulation vs. various temporal parameters for S-T events, including saccade-tail flip (d) onset delay, (e) offset delay, (f) overlapping time interval, (g) saccade duration, and (h) tail flip duration. (i) & (j) Scatter plots showing (i) tail flip asymmetry and (j) average undulation vs. stimulus delay for S-T events. In (d)–(j) Individual data points are shown in squares, diamonds and circles for the blank control, appetitive and aversive chemical groups respectively. Lines show best linear fits of each data group and shadows show 95% prediction confidence intervals of the blank control data group in (d) & (f) (the only that exhibited significant associations) and aversive chemical data group in (i) (the only that exhibited significant association). Pearson’s correlation coefficient (r) and associated P-value are shown for the (d) & (f) blank control and (i) aversive chemical data points.

Brain-wide chemical representation.

(a) Mean intensity (to transverse and sagittal planes) of mean mutual information (M.I.) between the calcium signals of appetitive valence-encoding regions-of-interest (ROIs) and the appetitive stimulus regressors ((Ivaline + Ialanine)/2, see Methods). (b) Regional mean ± S.E.M. (across larvae) of mean appetitive M.I. ((Ivaline + Ialanine)/2). P-value: Two-sided t-test comparing the pallium and the subpallium. (c) & (d) Similar to (a) & (b) but for mean aversive chemical M.I. ((Icadaverine + Iputrescine)/2) and aversive valence-encoding ROIs. (e) Mean intensity projections (to transverse and sagittal planes) of M.I. between the calcium signals of 1 mM valine-encoding ROIs and the stimulus regressor (Ivaline, see Methods). (f) Regional mean ± S.E.M. (across larvae) of sensory M.I. (g)–(l) Similar to (e) & (f) but for (g) & (h) alanine (Ialanine), (i) & (j) cadaverine (Icadaverine) and (k) & (l) putrescine (Iputrescine), all at 1 mM. The major brain regions (Tel: telencephalon, Di: diencephalon, Me: mesencephalon and Rh: rhombencephalon) are color-coded. Abbreviations of brain regions: Same as in Figure 5d. In all plots, data are pooled from n = 5 larvae. Scale bars: 50 µm in the Z-Brain Atlas space.

Additional analysis on brain-wide valence and motor representations.

(a) Upper panel: Pie charts showing the regional distributions of regions-of-interest (ROIs) (median proportions across larvae) with motor preference in different ranges. Lower panel: Mean ± S.E.M. (across larvae) responses of motor-encoding ROIs with motor preference in different ranges. Dotted vertical lines indicate the times of tail flip onsets (in a corresponding neuronal imaging frame). (b) Upper panel: Pie charts showing the regional distributions of ROIs (median proportions across larvae) with valence preference in different ranges. Lower panel: Mean ± S.E.M. (across larvae) responses of valence-encoding ROIs with valence preference in different ranges. Dotted vertical lines indicate the time points of stimulus onsets (between two neuronal imaging frames). In (a) and (b), the major brain regions (Tel: telencephalon, Di: diencephalon, Me: mesencephalon and Rh: rhombencephalon) are color-coded. The neural activity data correspond to the behavioral data reported in Figures 2 & 4, collected from n = 5 larvae. (c) & (d) Zoom-in maps showing the forebrain regions olfactory bulb, pallium and subpallium ROIs preferring (c) S-T event and aversive valence in Figure 5h, and (d) independent tail flip and appetitive valence in Figure 5j. Scale bars: 50 μm in the Z-Brain Atlas space.

Neuronal activity associated with chemosensory-driven coupled saccade-tail flip (S-T) events across different brain regions.

Mean ± S.E.M. (across n = 5 larvae) responses of motor-encoding regions-of-interest (ROIs) in the different brain regions, aligned to the onset of chemosensory cue-associated S-T events. Dotted traces show the corresponding neuronal responses of spontaneous S-T events. Dotted vertical lines indicate tail flip onsets (in a corresponding neuronal imaging frame). P-values: Right-sided Z-tests comparing pre-chemical cue-associated event neuronal activity at each time point to zero. Significance was set at 2.5 standard deviations (S.D.), with S.D. estimated from spontaneous event baseline activity variability. Adjustment for multiple comparisons across time points and brain regions was made using the Benjamini-Hochberg procedure.

Additional analysis on neural activity underlying chemosensory-driven coupled saccade-tail flip (S-T) events.

Mean ± S.E.M. (across n = 5 larvae) responses of motor-encoding regions-of-interest (ROIs) in the pallium, rhombomere 4 and rhombomere 6, classified according to (a) saccade-tail flip onset delay [< 0.1 seconds (upper) or ≥ 0.1 seconds (lower)] and (b) stimulus-tail flip delay [< 5 seconds (upper) or ≥ 5 seconds (lower)], and aligned to the onset of chemosensory cue-associated S-T events. Dotted traces show the corresponding neuronal responses of spontaneous S-T events (note that in (b), spontaneous S-T events were not classified according to stimulus-tail flip delay). Dotted vertical lines indicate tail flip onsets (in a corresponding neuronal imaging frame). P-values: Right-sided Z-tests comparing pre-chemical cue-associated event neuronal activity at each time point to zero. Significance was set at 2.5 standard deviations (S.D.), with S.D. estimated from spontaneous event baseline activity variability. Adjustment for multiple comparisons across time points and brain regions was made using the Benjamini-Hochberg procedure. Statistical significance: Black, significant; gray, near-significant.

Neuronal activity associated with chemosensory-driven independent tail flips (T events) across brain regions.

Mean ± S.E.M. (across n = 5 larvae) responses of motor-encoding regions-of-interest (ROIs) in the different brain regions, aligned to the onset of independent tail flips during chemical cue presentation. Dotted traces show the corresponding neuronal responses of spontaneous independent tail flips. Dotted vertical lines indicate tail flip onsets (in a corresponding neuronal imaging frame). P-values: Right-sided Z-tests comparing pre-chemical cue-associated event neuronal activity at each time point to zero. Significance was set at 2.5 standard deviations (S.D.), with S.D. estimated from spontaneous event baseline activity variability. Adjustment for multiple comparisons across time points and brain regions was made using the Benjamini-Hochberg procedure.