Evidence of negative affective state in Cavalier King Charles Spaniels with syringomyelia

Abstract

Syringomyelia is a common and chronic neurological disorder affecting Cavalier King Charles Spaniels. The condition is putatively painful, but evaluating the affective component of chronic pain in non-human animals is challenging. Here we employed two methods designed to assess animal affect – the judgement bias and reward loss sensitivity tests – to investigate whether Cavalier King Charles Spaniels with syringomyelia (exhibiting a fluid filled cavity (syrinx) in the spinal cord of ≥2mm diameter) were in a more negative affective state than those without the condition. Dogs with syringomyelia did not differ in age from those without the condition, but owners reported that they scratched more (P<0.05), in line with previous findings. They also showed a more negative judgement of ambiguous locations in the judgement bias task (P<0.05), indicating a more negative affective state, but did not show a greater sensitivity to loss of food rewards. These measures were unaffected by whether the dog was or was not receiving pain-relieving medication. Across all subjects, dogs whose owners reported high levels of scratching showed a positive judgement bias (P<0.05), indicating that scratching was not directly associated with a negative affective state. Tests of spontaneous behaviour (latency to jump up to or down from a 30cm high platform) and physiology (thermography of the eye) did not detect any differences. These results provide initial evidence from the judgement bias task that syringomyelia may be associated with negative affect in dogs, and open the way for further and larger studies to confirm findings and investigate the effects of medication in more detail

Impact of canine epilepsy on judgement and attention biases

“…Responses to the ambiguous stimuli in the judgement bias task differed from some other studies. Rather than an intermediate response to the middle bowl (as seen in Cockburn et al 37 and others), dogs showed similar response latencies to NP and MID, with a significant increase in latency to NN. Such response patterns were previously described in Muller et al 76 , and were speculated to be attributable to personality differences.…”

“…This is less likely, given our sample of over 35 dogs for each task (judgement bias n = 37; attention bias, n = 36), which is higher than other comparable studies that detected differences in affective state using the judgement bias paradigm e.g. eight dogs with syringomyelia vs. 13 controls 37 . However, as epilepsy is a heterogenous disease, even within-breed, it is possible that a larger sample is needed.…”

“…In addition, as this study tested the unknown affect treatment of epilepsy in a novel attention bias task, further validation of this task is required, including exploring the responses of dogs with known affect treatments e.g. syringomyelia in Cavalier King Charles Spaniels 37 , or separation related behaviour 38 , both of which have been found to induce a negative affective state in dogs, with this paradigm.…”

“…The judgement bias paradigm has been used to provide evidence of long-term impacts on affective state in dogs, including neurological disorders (e.g. syringomyelia in Cavalier King Charles Spaniels 37 ) and behavioural disorders (e.g. separation-related behaviours 38 ) with affected dogs exhibiting more negative judgements of the ambiguous stimuli than controls.…”

Using judgment bias test in pet and shelter dogs (Canis familiaris): Methodological and statistical caveats

“…Some studies in this area, whilst reporting a significant association between emotions and cognitive biases in dogs, have reported methodological and statistical issues, including a small sample size that makes it hard to infer a general behavioural pattern [ 37 , 38 ], the employment of a single ambiguous trial, the outcome of which could be influenced by momentary distraction [ 39 , 40 ] and the use of a statistical approach that evaluates the average dog’s response instead of single trial responses, thereby reducing variability in the data [ 27 , 31 , 41 ]. Averaged measures are inevitably less accurate since, during the test, trials are repeated for each dog and for each type of cue and the number of repetitions is not consistent among cues (more trials for each trained cue, less trials for each ambiguous cue to minimize a potential learning effect).…”