Developmental models for neurally based multisensory studies
One of the preeminent model species for studies of sensory (and multisensory) processes has been the cat. The reasons for this are manifold, but are partly a result of the marked visual, auditory, and tactile acuity of this species, and the striking commonalities in the organization of its sensory systems (and sensory brain regions) and those of primate species, including man.
Seminal studies into the operations performed by individual multisensory neurons were first carried out in the cat midbrain superior colliculus (SC). The cat SC was an ideal model for this work for a number of reasons, most notably because it was known as a site for sensory convergence from at least three different sensory systems (vision, hearing, and touch), because its stereotyped topographic organization made it very amenable to neurophysiological analyses, and because of the SC’s well-defined and well-characterized role in the orientation of gaze (see Stein and Meredith 1993).
Although the SC has remained an important model for multisensory research, more recent interest has focused on better detailing the organization and response characteristics of cortical multisensory domains, with the rationale that perceptual processes are more the domain of cortex as opposed to subcortex. In the cat, this work has focused on the cortex of the anterior ectosylvian sulcus (AES), an enigmatic cortical area comprised of three overlapping unisensory representations (visual, auditory, and somatosensory) whose role in perceptual processes remains to be elucidated (Stein and Wallace 1996; Wallace et al. 1992).
Although these studies in cat subcortical and cortical multisensory structures represent the foundation of work completed to date on questions of multisensory neural encoding (and as outlined below have provided a set of fundamental principles for multisensory processing), recent work has expanded the number of species investigated in multisensory research to include rodents, ferrets, and non-human primates (Allman et al. 2008; Avillac et al. 2007; Bizley et al. 2007; Foxe etal. 2002; Ghazanfar etal. 2005; Kayser etal. 2005; King and Palmer 1985; Wallace et al. 1996, 2004b). In each of these species, the basic operations performed by multisensory neurons and networks appear to be strikingly similar, suggesting a universal set of principles that govern the integration of sensory information. Briefly, multisensory integration in this context refers to the active process of evaluating and synthesizing information arriving from more than one sensory modality at the single-neuron level, a process that typically can dramatically transform the information encoded by that neuron (as evidenced by firing-rate changes under multisensory conditions).
From a developmental perspective, the cat has been the preferred species to date, not only for the reasons outlined above, but also because of its relatively short gestation time and the altricial (i.e. relatively immature) state of the newborn nervous system. Although most of the subsequent narrative will summarize developmental studies carried out in the cat model, it is important to point out that both rodents and non-human primates offer distinct advantages over cat (i.e. ease of developmental perturbation/manipulation and closer parallels to the human developmental processes, respectively) and are becoming increasingly important model species. Likewise, the ferret, another carnivore, is gaining popularity in the multisensory arena (Bizley et al. 2007; Keniston et al. 2009; Meredith and King 2004; Ramsay and Meredith 2004). Although lacking some of the foundational work needed to relate unisensory and multisensory systems, the ferret is much more readily trained in behavioural studies than the cat. Hence, future studies structured to more directly examine the correlative and causal links between multisensory networks and behavioural and perceptual processes may be more tractable in the ferret model.
