Investigating the potential of non-invasive seizure forecasting in canine epilepsy

Epilepsy is brain disease characterised by an enduring predisposition to spontaneously generate epileptic seizures, and affects ~0.6% of the canine population. Seizures are considered spontaneous due to our present lack of understanding of the mechanisms underlying the transition from interictal-ictal state. The ability to detect this transitional state (termed the ‘prodrome’), and forecast seizure activity would be beneficial to dog and owner alike. In people, behavioural and mood changes are widely reported during the prodrome, and there are anecdotal reports of analogous canine behavioural changes from owners; however, this has yet to be studied scientifically.

At present, intracranial-EEG is the only system that has been investigated in canine seizure forecasting, but is not available in the clinical context and is highly invasive. The aim of this PhD is to

1. assess the feasibility of non-invasive seizure forecasting for canine epilepsy using a combination of:

◦extra-cerebral biological signals: continuous logging inertia measurement units (IMUs) that capture movement information
◦heart rate monitors capable of measuring heart rate variability (R-R interval), and
◦ owner-reports of behavioural changes and seizure triggers

2. to identify internally and externally-derived seizure triggers, including physiological effects (e.g. sleep quality) and environmental effects (e.g. weather).

The PhD candidate will recruit dogs (n=30) with epilepsy (seizure frequency ≥3 seizures/month) from the RVC Small Animal Referral Hospital and first opinion practices into a longitudinal study, with recordings performed over 2 months at home. Algorithms will be developed to detect behavioural and physiological changes indicative of seizures, prodromal activity and sleep quality.

Methods to trial include:
•specific-behaviour detection;
•behavioural complexity measures;
•deviation from expected.

This project will yield a greater understanding of pre-seizure behaviour in dogs, and if algorithms are successfully developed, new treatment approaches can be explored including acute treatment interventions to abort impending seizures.

Essential requirements:

You do not need to be a qualified vet for this studentship but:
•You must have, or expect to achieve, a minimum of an upper second class Honours degree or international equivalent, in a subject related to behavioural sciences, mathematics or computer science.

Desirable requirements:
•If your degree is in behavioural sciences you must demonstrate an interest in computer science or mathematics.
•If your degree is in mathematics or computer science, you must demonstrate an interest in animal behaviour or welfare.