Abstracts

Rationale: Zebrafish larvae have emerged as efficient tools in translational epilepsy research (Baraban et al. 2005). We recently developed methods for noninvasive brain electrical recordings from zebrafish larvae with the use of a 61-channel multi-electrode array (MEA). As we are recording from multiple electrodes, we asked whether excessive synchrony of neuronal firing, a hallmark of epileptic seizures, can be detected with our novel methods for epicranial seizure recording in zebrafish larvae.Methods: We recorded action potentials from intact larval zebrafish brains at 4- to 5-days post fertilization; the larvae were restrained on an MEA during continuous flow of artificial cerebrospinal fluid (aCSF). Seizures were induced by the addition of 15 mM potassium chloride (KCl; n=6) or 15 mM pentylenetetrazole (PTZ; n=6) added to the perfusion fluid. Average spike (action potential) frequency and average spike burst count (per 1800 secs) before and after the addition of the pro-convulsant were computed for all channels. Bursts were defined as time-limited increases in spike rates in excess of twice the standard deviation of the average spike rate per channel. Spatial coherence was obtained by computing the per-second r-value between the spike rate of an individual channel and the average rate of neighboring channels. Temporal clustering was computed to determine the ratio of the number of spikes within defined clusters (epochs of continuous firing in excess of 1 Hz) to the total spikes across 1800 secs.Results: Extracellular action potentials were reliably recorded and were stable over time and across experimental conditions. From most channels (90 %, 692 / 767 channels), we recorded units with baseline spontaneous bursting activity and a smaller proportion (10 %, 75 out of 767 channels) of units that fired continuously at baseline. Burst rate increased after the addition of KCl (p = 0.019) or PTZ (p = 0.032), as did spike rate (p = 0.023 for KCl, p = 0.013 for PTZ). In the control condition, in the absence of either KCl or PTZ, neither an increase of burst count nor spike rate was recorded. Relative to baseline, post PTZ, the spiking patterns were spatially coherent (p<0.05) and temporally clustered (p<0.05). At time of this writing, analysis of KCl-mediated changes in spatial coherence and temporal clustering is in progress. Additionally, in 14 experiments, larvae were monitored for long-term survival following a minimum of a two-hour recording session during which no drug was added. 12/14 survived at least for 24 hours, and 7/14 for more than 48 hours (maximum 10 days). In one fish, we obtained high-quality recordings on consecutive days.Conclusions: Seizures in larval zebrafish are characterized by increased neuronal firing rates, clustering of firing patterns over time, and an increase in the synchrony of neuronal firing patterns across brain regions. An advantage of our protocol is also fish survival, which enables longitudinal and recurrent fish monitoring.