Objective

Ictal vocalizations have shown diagnostic utility in epilepsy patients. Audio recordings of seizures have also been used for seizure detection. The present study aimed to determine whether generalized tonic-clonic seizures in the Scn1a^+/- mouse model of Dravet syndrome are associated with either audible mouse squeaks or ultrasonic vocalizations.

Methods

Acoustic recordings were captured from group-housed Scn1a^+/- mice undergoing video-monitoring to quantify spontaneous seizure frequency. We generated audio clips (n = 129) during a generalized tonic-clonic seizure (GTCS) that included 30 seconds immediately prior to the GTCS (preictal) and 30 seconds following the conclusion of the seizure (postictal). Nonseizure clips (n = 129) were also exported from the acoustic recordings. A blinded reviewer manually reviewed the audio clips, and vocalizations were identified as either an audible (<20 kHz) mouse squeak or ultrasonic (>20 kHz).

Results

Spontaneous GTCS in Scn1a^+/- mice were associated with a significantly higher number of total vocalizations. The number of audible mouse squeaks was significantly greater with GTCS activity. Nearly all (98%) the seizure clips contained ultrasonic vocalizations, whereas ultrasonic vocalizations were present in only 57% of nonseizure clips. The ultrasonic vocalizations emitted in the seizure clips were at a significantly higher frequency and were nearly twice as long in duration as those emitted in the nonseizure clips. Audible mouse squeaks were primarily emitted during the preictal phase. The greatest number of ultrasonic vocalizations was detected during the ictal phase.

Significance

Our study shows that ictal vocalizations are exhibited by Scn1a^+/- mice. Quantitative audio analysis could be developed as a seizure detection tool for the Scn1a^+/- mouse model of Dravet syndrome.