The innate ability of pet dogs to spontaneously anticipate epileptic seizures

The main aim of this Study was to investigate anecdotal claims of spontaneous seizure alerting behaviour by pet dogs and then to attempt to elucidate the sensory mechanism they used for doing so. Little was known about this topic when I started my research in 2014, when, to the best of my knowledge, only four limited studies had been published. Chapter one defines epilepsy as a serious neurological condition which is often associated with injury and comorbidities like anxiety, depression, suicidal intention, limited employment prospects, social isolation and prejudice. In global terms, thirty per cent of people with epilepsy, (20 million), are unable to control their seizures by medication. For them, the only intervention available is targeted brain surgery to excise the seizure source, but waiting lists are long and the procedure expensive. Anecdotally, dogs appear to be capable of anticipating seizures although no empirical evidence has been offered to support this claim. Many accounts in the literature, attest to the role dogs have played in assisting humanity since they were first domesticated some 30,000 years ago. They have shown an astonishing ability to detect concealed contraband in support of law enforcement, or to find people trapped in hostile terrain or under collapsed buildings. Dogs have also begun to assist with medical diagnosis of life-threatening conditions like cancer and in the prediction of hypo and hyper glycaemic events albeit with questionable reliability. Chapter two recounts the outcome of an international survey of people with epilepsy, 70% of whom attested their pet dogs had demonstrated spontaneous seizure-alerting. It appears this is a widespread phenomenon, not restricted by nationality or social strata and with a commonality of behavioural responses resembling attention-seeking strategies. These seizure-related behavioural changes were found to be independent of breed, gender or length of ownership, but were predicted by the absence of children and by the relationship between owners and dogs. Furthermore, alerting responses appeared to be consistent and reliable for most or in some cases, all seizure events, with predictive times ranging from minutes to hours prior to an episode. Chapter three explored the viability and reliability of a bespoke piece of apparatus called a remote odour delivery mechanism, (RODM), for transferring odours from A to B. This was tested by delivering target odours to six specialist search dogs to compare their trained target-response behaviours with their reaction to the RODM delivered target odours. Chapter Four describes how the RODM facilitated the controlled observation of responses made by nineteen untrained dogs to seizure odours, delivered to the vicinity of their owners. The subsequent reactions of the dogs strongly reflected those reported as being of an attention-seeking nature, described for spontaneous seizure alerting in Chapter Two. These findings provided compelling evidence for the existence of a seizure-related olfactory bio-marker. To attempt to understand why 30% of the dogs sampled in Chapter Two did not respond to seizure onset, Chapter five explored whether this could be a factor of the dogs’ relationships with their owners. This relationship and some traits of the dogs, were assessed using two validated owner-dog assessment mechanisms, the findings of which indicated that three characteristics were key predictors for seizure alerting behaviour: (1) an enhanced relationship occasioned by an active owner-dog association, (2) low scores on separation anxiety and (3) low scores on trainability.