What could evolve in the evolution of memory?

Abstract

Over the past five decades, advances in neuroscience have set out the adaptive landscape of memory, illuminating semi-independent storage mechanisms, forgetting mechanisms and modifications to basic machinery that bring context specificity. Yet because much of neuroscience aims to understand how the human brain functions, rather than to explore taxonomic diversity, the implications for animal cognitive evolution remain underexplored. This perspective article examines the potential evolutionary diversity of animal memory from a mechanistic viewpoint. We argue that taking into account neurogenetic and neurophysiological mechanisms of memory could illuminate how the diversity of cognitive traits has been shaped by natural selection. By focussing on memory in insects, notably through the lens of associative processes, we target our discussion on potential variation in taxonomically general processes within one of the animal kingdom's richest and most diverse animal groups. This exploration aims to broaden the discourse on memory evolution within the field of cognitive ecology towards an understanding of the many ways in which memory could be shaped by natural selection.This article is part of the Theo Murphy meeting issue 'Selection shapes diverse animal minds'.

Keywords: behaviour; cognition; evolution; memory.

Figure 1.

Models of memory phases in (a) honeybees (Apis mellifera; appetitive conditioning) and (b) fruit flies (Drosophila melanogaster; aversive conditioning). Memory is generally divided into short-term (STM), intermediate-term (ITM, also referred to as middle-term MTM) and long-term (LTM) memory phases, although the timings for each phase vary between species and protocols. Memory in honeybees is generally measured using appetitive conditioning protocols, and LTM is further divided into an early (eLTM) and a late (lLTM) phase, which are characterized by translation only, or transcription and translation, respectively. By contrast, memory in fruit flies is generally measured using aversive conditioning protocols. Flies can form a type of long-term memory (LTM) that is not dependent on protein synthesis (anaesthesia-resistant memory, ARM), and a late-phase long-term memory (LP-LTM) that is protein synthesis-dependent. Honeybee figure based on [14]; fly figure based on [8,9,36].

Figure 2.

A model of aversive olfactory conditioning in D. melanogaster. (a) Simultaneous or temporally overlapping exposure to an odour (conditioned stimulus; CS) and electric shock (unconditioned stimulus; US) results in learnt avoidance of that odour. (b) Exposure to the odour activates a unique subset of mushroom body neurons (MB neurons), which constitute the neural trace of the CS. Dopaminergic neurons (DANs) from the PPL1 cluster synapse into the MB neurons and respond to the negative valence of the US. (c) Coactivation of any MB neuron and an incoming PPL1 DAN leads to suppression of nearby synapses with mushroom body output neurons (MBONs). (d) The mushroom bodies are divided into discrete compartments, such that incoming negative-valence PPL1 DANs are close to approach-driving MBONs, and positive-valence PAM DANs are close to their avoidance driving equivalents (note that ’approach-driving’ and ’avoidance-driving’ are simplifications of a much more complex reality, for the purpose of illustration). Aversive input therefore suppresses only approach behaviour and appetitive input suppresses avoidance behaviour. Figure is based on model description in [9,53,54]