Spatial learning through active electroreception in Gnathonemus petersii

Mormyrid fish as models for investigating sensory-motor integration: A behavioural perspective

Animals possess senses which gather information from their environment. They can tune into important aspects of this information and decide on the most appropriate response, requiring coordination of their sensory and motor systems. This interaction is bidirectional. Animals can actively shape their perception with self-driven motion, altering sensory flow to maximise the environmental information they are able to extract. Mormyrid fish are excellent candidates for studying sensory-motor interactions, because they possess a unique sensory system (the active electric sense) and exhibit notable behaviours that seem to be associated with electrosensing. This review will take a behavioural approach to unpicking this relationship, using active electrolocation as an example where body movements and sensing capabilities are highly related and can be assessed in tandem. Active electrolocation is the process where individuals will generate and detect low-voltage electric fields to locate and recognise nearby objects. We will focus on research in the mormyrid Gnathonemus petersii (G. petersii), given the extensive study of this species, particularly its object recognition abilities. By studying object detection and recognition, we can assess the potential benefits of self-driven movements to enhance selection of biologically relevant information. Finally, these findings are highly relevant to understanding the involvement of movement in shaping the sensory experience of animals that use other sensory modalities. Understanding the overlap between sensory and motor systems will give insight into how different species have become adapted to their environments.

The Use of Supervised Learning Models in Studying Agonistic Behavior and Communication in Weakly Electric Fish

Despite considerable advances, studying electrocommunication of weakly electric fish, particularly in pulse-type species, is challenging as very short signal epochs at variable intervals from a few hertz up to more than 100 Hz need to be assigned to individuals. In this study, we show that supervised learning approaches offer a promising tool to automate or semiautomate the workflow, and thereby allowing the analysis of much longer episodes of behavior in a reasonable amount of time. We provide a detailed workflow mainly based on open resource software. We demonstrate the usefulness by applying the approach to the analysis of dyadic interactions of Gnathonemus petersii. Coupling of the proposed methods with a boundary element modeling approach, we are thereby able to model the information gained and provided during agonistic encounters. The data indicate that the passive electrosensory input, in particular, provides sufficient information to localize a contender during the pre-contest phase, fish did not use or rely on the theoretically also available sensory information of the contest outcome-determining size difference between contenders before engaging in agonistic behavior.

Linking active sensing and spatial learning in weakly electric fish

Weakly electric fish can learn the spatial layout of their environment using only their short-range electric sense. During spatial learning, active sensing motions are used to memorize landmark locations so that they can serve as anchors for idiothetic-based navigation. A hindbrain feedback circuit selectively amplifies the electrosensory input arising from these motions. The ascending electrolocation pathway preferentially transmits this information to the pallial regions involved in spatial learning and navigation. Similarities in both behavioral patterns and hindbrain circuitry of gymnotiform and mormyrid fish, two families that independently evolved their electrosense, suggest that amplification and transmission of active sensing motion inputs are fundamental mechanisms for spatial memory acquisition.

Neural Networks: How a Multi-Layer Network Learns to Disentangle Exogenous from Self-Generated Signals

Artificial multi-layer networks can learn difficult tasks, such as recognizing faces, but their architecture and learning rules appear to be very different from those of biological neural networks. Experimental and computational studies of a two-layered biological neural network have revealed how the learning rules used in artificial neural networks can be efficiently implemented by neurons with complex dynamics and precisely organized connectivity.