From representations to servomechanisms to oscillators: my journey in the study of cognition

Abstract

The study of comparative cognition bloomed in the 1970s and 1980s with a focus on representations in the heads of animals that undergird what animals can achieve. Even in action-packed domains such as navigation and spatial cognition, a focus on representations prevailed. In the 1990s, I suggested a conception of navigation in terms of navigational servomechanisms. A servomechanism can be said to aim for a goal, with deviations from the goal-directed path registering as an error. The error drives action to reduce the error in a negative-feedback loop. This loop, with the action reducing the very signal that drove action in the first place, is key to defining a servomechanism. Even though actions are crucial components of servomechanisms, my focus was on the representational component that encodes signals and evaluates errors. Recently, I modified and amplified this view in claiming that, in navigation, servomechanisms operate by modulating the performance of oscillators, endogenous units that produce periodic action. The pattern is found from bacteria travelling micrometres to sea turtles travelling thousands of kilometres. This pattern of servomechanisms working with oscillators is found in other realms of cognition and of life. I think that oscillators provide an effective way to organise an organism’s own activities while servomechanisms provide an effective means to adjust to the organism’s environment, including that of its own body.

Photobehaviours guided by simple photoreceptor systems

“…And third are servomechanisms, behavioural adjustments made to compensate or correct for a perceived error on the animal’s current course of behaviour (Cheng 2022 ). Many locomotive actions such as walking or swimming are controlled with rhythmic behaviours: oscillators and the input from photoreceptors can aid in navigation via servomechanism systems, a feedback loop in which the action being carried out serves to reduce the divergence from the sensory system’s ideal (Cheng 2023 ). An example of a reflex might be the aversive contraction responses of fan worms to passing shadows, whereas most navigational tasks like phototaxis are controlled with servomechanisms, corrections on directional course in response to changes in sensory cues such as light intensity.…”

Active oscillations in microscale navigation

“… 2022 ) This raises the possibility that the ancestral analogue of central pattern generators (CPGs)—ubiquitous in animal locomotion Marder and Bucher ( 2001 )—may have arisen in the first instance in the single-celled eukaryote, purely from the necessity of coordinating multiple oscillators for motility and navigation (Wan 2020 ). These basic mechanisms are then elaborated in the complex nervous system architectures of animals, where intrinsically oscillatory appendage actuation obeys different network topologies, and regulation by servomotors (Cheng 2022 , 2023 ). CPG-inspired circuits also underlie the many designs for robots modelled on the locomotor patterns of animals Ijspeert ( 2008 ).…”

Directed retreat and navigational mechanisms in trail following Formica obscuripes

“…Extensive research by Ken Cheng and colleagues on navigation in ants has highlighted the value of a comparative ecological approach in which mechanisms are studied in the context of their function in the natural environment (e.g., Bühlmann et al, 2011 , 2020 ; Cheng et al, 2009 , 2012 , 2014 ; Cheng, 2022 , 2023 ; Freas & Cheng, 2022 ; Freas et al, 2018 , 2019a , b , c ; Schultheiss et al, 2016 ). Comparisons between ant species have revealed many similarities that exist across a wide range of environments and foraging ecologies, with the presence of a common underlying navigational toolkit of concurrently operating strategies (Bühlmann et al, 2011 ; Cheng et al, 2009 ; Freas & Spetch, 2023 ; Wehner, 2020 ).…”