Choosing sides: making decisions in an escape response

Many of our movements are made without great deliberation. For instance, removing a hand from a hot stove is an easy decision to make. But what determines whether we choose to start walking with our right foot rather than our left? (Briggman et al. 2005). The decision-making processes that underlie our movements, the why and how we move the way we do, are influenced by many factors such as what our senses tell us about our bodies and the external world, our motivations, and considerations of costs based upon energetics, time, distance, perceived success rates and outcomes. Our decisions are further shaped by uncertainty associated with our perception of sensory information and variability in how our bodies perform movements (Wolpert & Landy, 2012). Considering the complexity of making motor decisions, simpler organisms such as invertebrates or vertebrates at early developmental stages have been used as model systems. In these systems, the neural architecture underlying the selection and execution of stereotypical behaviours such as swimming, feeding, or mating can be mapped to a finite number of well-described neurons. Neural control of these movements, including the selection of different motor programmes, can be more readily studied. In this issue of The Journal of Physiology, Buhl and colleagues take advantage of the relatively well-described circuitry underlying the control of swimming in the Xenopus tadpole to uncover a novel mechanism by which the decision as to which side to initiate an escape response is made. Across species, behaviours or motor programmes seem to be under the control of command neurons or decision-makers. Each decision-maker is associated with a particular motor programme and their activity determines which motor programme is selected. These decision-makers integrate over time excitation derived from sensory information. In a competitive process, the first decision-maker to accumulate enough evidence to reach a certain threshold initiates its associated motor programme and other motor programmes are inhibited. This ‘ramp-to-threshold’ behaviour can be implemented by single neurons (Murakami et al. 2014) or by a population of neurons acting as decision-making kernels (Briggman et al. 2005). These decision-making kernels may consist of ramp-to-threshold neurons as well as other neurons involved in setting the dynamics of the activity of the kernel as a whole (Murakami et al. 2014). In the Xenopus tadpole, a class of reticulospinal neurons termed descending interneurons (dINs) play the role of decision-makers when it comes to initiating swimming movements (Soffe et al. 2009). Buhl and colleagues investigated their roles in deciding the direction of an escape response following a touch to the head. In response to a tactile stimulus applied to the head, tadpoles will react with a body flexion to either side followed by swimming away in an unpredictable direction. The decision whether to flex towards or away from the stimulated side is more variable in response to weaker stimuli to the head. Their experiments reveal that the key to this decision lies within an asymmetry in the manner that sensory stimulation is communicated to dINs on the stimulated and the unstimulated side of the body. In immobilized preparations, dINs on the stimulated side were found to receive low-latency, fast rising excitation, whereas dINs on the opposite side received longer-latency but steadily growing excitation. This longer-latency excitation would reliably induce a firing burst on the opposite side at a set timing after stimulation that would lead to flexion away from the stimulus. The decision of flexing towards or away from the head touch rests upon whether the early-rising excitation to the dINs on the stimulated side is sufficiently reliable to induce a first burst on this side, which would precede the reliable longer-latency burst on the unstimulated side. Further electrophysiological recordings, in combination with lesioning experiments, showed that the early-rising excitation to dINs on the stimulated side originates from neurons in the trigeminal nucleus (tINs) that are excited by trigeminal sensory neurons responding to head touch (Buhl et al. 2012). A high failure rate of synaptic transmission between tINs and dINs explains the stochastic appearance of the first burst on the stimulated side. On the unstimulated side, dINs reliably receive longer-latency excitation via a polysynaptic pathway involving a newly identified population of commissurally projecting dorsolateral interneurons. This study therefore reveals that the decision to flex to either side of the body following a head touch is based upon a race towards a threshold (the threshold to action potential firing) between two populations of decision makers, dINs on the stimulated side versus dINs on the unstimulated side. Stochasticity in this decision is a result of asymmetries of neural connectivity and synaptic transmission in the pathways linking tactile stimuli to the head and the two populations of dINs on either side of the body. A beneficial feature of this circuitry is that its asymmetry introduces appropriate delays in synaptic transmission that preclude the co-contraction of the stimulated and unstimulated side that would prevent an effective escape manoeuvre to be executed. Buhl et al. suggest that stochasticity in the escape response of tadpoles may be a strategy to prevent predators from taking advantage of a predictable behaviour. Their study reveals that this stochasticity is a consequence of asymmetries in neural connectivity and synaptic transmission involving the decision-maker dINs on both sides of the body. By biasing the escape response at early developmental stages to exhibit motor noise, tadpoles are perhaps preventing the competition between prey and predator from being a one-sided affair. None declared.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.