Abstract:Neurons in the brain have rich and adaptive input-output properties. Features such as heterogeneous f-I curves and spike frequency adaptation are known to place single neurons in optimal coding regimes when facing changing stimuli. Yet, it is still unclear how brain circuits exploit single-neuron flexibility, and how network-level requirements may have shaped such cellular function. To answer this question, a multi-scaled approach is needed where the computations of single neurons and neural circuits must be considered as a complete system. In this work, we use artificial neural networks to systematically investigate single-neuron input-output adaptive mechanisms, optimized in an end-to-end fashion. Throughout the optimization process, each neuron has the liberty to modify its nonlinear activation function parametrized to mimic f-I curves of biological neurons, either by learning an individual static function or via a learned and shared adaptation mechanism to modify activation functions in real-time during a task. We find that such adaptive networks show much-improved robustness to noise and changes in input statistics. Using tools from dynamical systems theory, we analyze the role of these emergent single-neuron properties and argue that neural diversity and adaptation play an active regularization role, enabling neural circuits to optimally propagate information across time. Finally, we outline similarities between these optimized solutions and known coding strategies found in biological neurons, such as gain scaling and fractional order differentiation/integration. Evolution has shaped neural circuits in the brain to support complex tasks and behavior. In doing so, single neurons have developed intriguing coding properties such as heterogeneous and adaptive nonlinear responses, and mechanisms at network levels have emerged to efficiently propagate and process information. Discerning the impact of these distinct scales in shaping biophysical mechanisms is a difficult proposition. Are circuits wired the way they are because of single-neuron properties? Are single neurons behaving the way they do because of circuit requirements? To tackle this question, we turn to neural network models that we optimize with deep learning techniques. Crucially, we endow single neurons in these models with parameters that govern their input-output properties, as well as adaptation dynamics. These mechanisms are then optimized end-to-end together with network parameters to perform various perception-like tasks. While this process could have produced any novel solution, we find that what naturally emerges from optimization bears noteworthy similarity to biophysical mechanisms found in neurons in the brain, such as spike-frequency adaptation and gain scaling. We perform an in-depth analysis of the optimization process that uncovered these strategies and formulate hypotheses relating joint evolutionary pressures of single-neuron properties and circuit-level requirements in the brain.

Evolution leads to a diversity of motion-detection neuronal circuits

Evolution of neural circuitry and cognition

Circuit design in biology and machine learning. I. Random networks and dimensional reduction

A visual motion detection circuit suggested by Drosophila connectomics

Developmental and evolutionary constraints on olfactory circuit selection

Evolutionary emergence of biological intelligence

Generation of motor program diversity and variability through inhibitory circuit motifs in the larval locomotor system

Fate of Duplicated Neural Structures

Neural networks with optimized single-neuron adaptation uncover biologically plausible regularization

Evolution of a central dopamine circuit underlies adaptation of light-evoked sensorimotor response in the blind cavefish, Astyanax mexicanus.

Environmental complexity and regularity shape the evolution of cognition

Normative and mechanistic model of an adaptive circuit for efficient encoding and feature extraction

Preserved neural dynamics across animals performing similar behaviour

Adapting to time: Why nature may have evolved a diverse set of neurons

Neural Elements for Predictive Coding

Explaining the Neuroevolution of Fighting Creatures Through Virtual fMRI

A Behavioural Perspective on the Early Evolution of Nervous Systems: A Computational Model of Excitable Myoepithelia

Circuit design in biology and machine learning. II. Anomaly detection

From Bayes to Darwin: evolutionary search as an exaptation from sampling-based Bayesian inference

On the design of neural networks in the brain by genetic evolution.

Evolutionary Implications of Self-Assembling Cybernetic Materials with Collective Problem-Solving Intelligence at Multiple Scales