temporal window for perception of simultaneity (Zampini et al. 2005). Therefore judgements about the temporal order of stimuli are more accurate when they are spatially separated than when they are in the same position. However, differences in speed and temporal processing mean that auditory and visual signals from the same object are not received at the same time away from the horizon of simultaneity and therefore have to be combined into a single multisensory object resulting in the loss of the information on the relative temporal onsets of the different stimuli (Spence et al. 2003).

2.1.4 Multisensory Perception in the Superior Colliculus

Convergence of visual, auditory and somatosensory (tactile and relating to the internal organs) stimuli occurs at several places in the central nervous system, including the superior colliculus, which is discussed in more detail in Section 2.3.6. Visual, somatosensory and auditory representations in the superior colliculus have similar map-like representations and therefore eye movements due to non-visual stimuli are directed to the source of the stimuli, as in the case of visually evoked eye movements. However, they are generally less accurate, as auditory receptive fields are larger than visual ones.

Activation of a particular region of the superior colliculus produces movements of the eyes, ears and head to focus on the area in space represented by the stimulation site. Therefore, the motor circuitry of the superior colliculus has a map-like organisation. The superior colliculus has a number of different layers. The organisation of both visual and somatopic stimuli is map-like with a spatial relationship between the location of the stimulus and a point on the superior colliculus. There is a very regular relationship between the visual and somatotopic maps in the deep layers, with the visual map emphasising the central retina, which has the greatest visual sensitivity, and the somatosensory map emphasising the face and hands, which have the greatest tactile sensitivity. There is also an auditory map, but it is tonotopic, that is, based on sound frequencies or tones, rather than the location of the stimulus, as with the visual and somatotopic maps. Therefore, the spatial auditory map uses differences between the intensity and timing of sounds at the two ears. This is because visual and somatosensory peripheral nerve fibres respond to a stimulus in a restricted area, regardless of its intensity, whereas auditory nerve fibres respond to loud sounds regardless of their spatial location. The spatial auditory map is oriented similarly to the visual and somatosensory maps.

The visual, auditory and somatosensory representations in the superior colliculus use common axes and the auditory and somatosensory components of the multisensory representation can be described in visual coordinates and vice versa. Other than a general increase in receptive field size with the number of sensory modalities, there are few differences between unimodal, bimodal and trimodal neurons. Therefore their large receptive fields mean that the receptive fields of a visual-auditory neuron are rarely completely coincident and the three receptive fields of a trimodal neuron are almost never coincident. This lack of precision