• calcium imaging
• epilepsy
• LFP
• local field potential
• zebrafish

Epilepsy is a neurological disorder characterized by recurrent spontaneous burst of synchronized activity, known as seizure. Furthermore, epilepsy is often accompanied by cognitive, behavioural, and psychiatric comorbidities that reduce the quality of life. Over the past decade, zebrafish (Danio rerio) has emerged as an important model organism for human disorders, as it represents an excellent compromise between system complexity and experimental accessibility. Its external embryonic development, high fecundity rate, small size and fully sequenced genome make the zebrafish easy to manage and manipulate. Indeed, in recent time, zebrafish has become a model organism for biomedical research for drug screenings but also to understand epilepsy mechanism. Calcium imaging provides information about the neuronal activity at low temporal frequency and high spatial resolution; LFP, on the contrary, is an integrated signal with a temporal resolution in the order of milliseconds. Given the different type of data provided by these two approaches, seemed to us crucial to combine them in order to reach an appropriate accuracy on both dimensions, spatial and temporal. In the present study, we combined LFP recording with two-photons Ca2+ imaging for the evaluation of two different zebrafish larvae models. In particular, we used the well-established pharmacological induced epileptic model upon exposure of the proconvulsant GABAa ligand pentylenetetrazol (PTZ) and a genetic model of the EAST/SeSAME syndrome knocking-down with the morpholino technology the mRNA of the kcj10a gene, coding for a family of potassium channels. For the critical assessment of the brain activity we first develop ‘ImageGateway’, a custom-made program written in MATLAB for the image analysis. This computational tool exploits the statistic of the imaging recordings in order to extract the relevant calcium fluctuations. The events we found form coherent spatio-temporal patterns that we were able to clusterize using a volumetric strategy and to extract the relevant features such as volume, duration, surface occupied and localization. We showed that those clusters can describe the peculiar dynamic of the two epileptic models and the correlation with the LFP. Then we assessed that all the extracted features follow a power-law distribution. Finally, we analysed the activity of single neurons providing general maps showing the neurons that follow the electrophysiological signal and neurons that precede that epileptiform events.