Overview

Assays and exposure

Twenty adult zebrafish were assigned to CTR, VPA (200 μM valproic acid), AO (250 μg/L Acmella oleracea extract), or VPA + AO, with five fish per group. Waterborne exposure lasted 48 hours. Behaviour was assessed 24 hours after the last administration. Each test used a 30-second accommodation period followed by a 300-second observation trial. The NTT used a 6 L glass tank. The simulation clock represents only the observation trial, not accommodation.

The NTT is shown in lateral view: the upper and lower halves represent vertical exploration. The MBT uses a top-down T-maze with the mirror at the left end. The symbolic fish and heading marker show simulated position and direction; the aquarium is a schematic, not video or a measured three-dimensional apparatus.

How the reconstruction works

A seeded, correlated random walk uses velocity and zone preferences derived from the approximate group means. NTT movement is restricted to the lower half until the modelled first-entry latency; illustrative pauses are parameterised by freezing time. MBT zone means are treated as relative preference weights because they do not consistently sum to 300 seconds. Mirror contacts are stochastic illustrative events, not detected bites.

Assumed schematic dimensions are 30 × 20 cm for NTT and 30 × 30 cm for MBT. These are model settings, not measured apparatus dimensions. The summary is calculated from the generated positions at 0.25-second intervals; therefore it need not reproduce the input means. The scaling factor linearly changes treatment deviations from CTR; it is neither a pharmacological prediction nor evidence of efficacy. CTR is unchanged by this factor.

Model limitations: this random walk is not fitted or validated against individual tracking records. In NTT, activity time, crossing count and total distance are outputs rather than enforced targets; model velocity, freezing and zone preferences influence the motion. A displayed group mean is not the expected result of an individual run. The model's average speed is path length divided by elapsed simulated time; it is not a replication of the original EthoVision velocity calculation. MBT activity and source velocity, maximum acceleration and rotation counts are available as source estimates but are not modelled as comparable outcomes. Synthetic mirror events are not observed bites.

Using the same seed, assay, group, and factor reproduces a trace; change the seed to generate another realisation. Generate trajectory calculates positions incrementally at 8× playback speed: 300 trial seconds take 37.5 seconds on screen. The time scaling does not change simulated velocities or distances. The trace, heatmap and summary accumulate together. Pause and Resume preserve the current trial. Static complete trace is optional. Leaving this section or hiding the tab pauses generation. PNG exports contain only the generated duration.

Data provenance and statistical interpretation

Bars are rounded visual estimates checked against the original plots, not exact author-supplied values or a calibrated digitisation. Small numerical differences must not be interpreted as new findings. Original analysis used one-way ANOVA with Tukey’s HSD: * p < 0.05, ** p < 0.01. Numerical SEM values were not extracted; the original error bars remain visible in Overview. No raw samples, exact p-values or new statistical tests are inferred. Brackets follow the explicit pairwise statements in Results. Each identified conflict with graphical annotations is displayed alongside the affected comparison and included in its export caption. Absence of a bracket means no significant comparison is encoded here, not proof of equivalence.

The parameter browser covers the distinct endpoints visible in the original plots. Some source legends enumerate additional inactivity panels that are not actually present; no separate inactivity values have been invented. Freezing is retained as duration in seconds, matching the plotted axis, rather than relabelled as frequency.

Proposed pathway · hypothesis, not a measured mechanism

Dashed arrows indicate hypotheses. Neurotransmitter levels and molecular signalling were not measured; no causal or clinical inference is established.