The superior colliculus (SC) is a sensory integration hub in the dorsal brainstem where multimodal information is combined and, depending on the saliency of the competing sensory inputs, appropriate motor commands and supportive autonomic changes initiated. In rodents, the SC is indispensable for initiating behavioural responses to stereotypical visual stimuli that resemble approaching objects, such as looming (an expanding overhead black circle). In pilot experiments we found that presentation of overhead looming stimuli drove acute surges in blood pressure in telemetered conscious rats. Surprisingly, equivalent autonomic responses could be driven by more complex visual stimuli (e.g. snakes). We hypothesized that the encoding capabilities of the SC might extend beyond detection of stereotypical approach cues conventionally attributed to the region, and that complex naturalistic shapes could be detected by SC circuits based on a saliency-map of behaviourally-relevant cues important for survival (from appetitive to threatening). To investigate this idea we made ensemble recordings from populations of SC neurons in anaesthetised rats and used machine-learning based approaches to decode the visual cues presented, clustering stimuli according to the similarity of population responses. We report categorical and invariant visual object capabilities of SC microcircuits that differ by subregion and are several orders of magnitude more complex than previously recognised. Our data suggest that the SC is capable of nuanced object recognition, suggesting mechanisms through which complex visual cues can initiate defence manoeuvres.
|Number of pages||2|
|Publication status||Published - 2 Dec 2019|
|Event||Australasian Neuroscience Society Annual Conference 2019 - Adelaide Convention Centre, Adelaide, SA, Australia|
Duration: 2 Dec 2019 → 5 Jan 2020
|Conference||Australasian Neuroscience Society Annual Conference 2019|
|Abbreviated title||ANS 2019|
|Period||2/12/19 → 5/01/20|