Good vibrations switch attention: An affective function for network oscillations in evolutionary simulations

Abstract

The method of computationally simulating evolutionary processes provides a unique opportunity for the automated development of models and hypotheses on cognitive and affective processes and their underlying neural mechanisms. The role of the modeler is limited to setting up the evolutionary selection procedure (e.g., a genetic algorithm; Holland, 1975), the initial state, and the environmental conditions. This has several advantages, of which the opportunity for innovation is not the least. Novel models with mechanisms and functions that have not been previously considered may emerge from these undirected optimization procedures.In previous simulation work (Heerebout & Phaf, 2010), we serendipitously discovered "artificial neural oscillations" in agents that inhabited a virtual environment (see Figure 1). The oscillations were an emergent property, not intentionally built in, and not even thought of beforehand, which proved to be highly adaptive. These simulations extended our investigation of LeDoux's (1996) evolutionary justification of his dual-pathway model for the processing of emotional stimuli (den Dulk, Heerebout, & Phaf, 2003). The agents, which were controlled by artificial neural networks, increased their chances of reproduction and survival by collecting food while avoiding predators. Analyses of the oscillating agents showed that the evolutionary advantage was conferred by an enhanced capacity for attentional switching when in an oscillatory mode.To test the switching efficacy of an oscillating agent, it was compared with a nonoscillating agent from a control simulation with a simpler network. Both agents responded to a plant that was placed in front of them at an angle of 45º to its left, which would, as the agent approached the plant, suddenly be replaced by a predator. The agents' phenotypic behavior was described in terms of speed of movement and rotational speed, and revealed distinct approach and avoidance behaviors. The oscillating agent would first "cautiously" approach the plant following a swerving, "zig-zag" trajectory, and then, as it detected the predator, would make a sharp turn (at 0.042º per time step) and accelerate strongly (the speed increased 87% in the first 10 time steps). The nonoscillatory agent kept a more constant speed. When it detected a predator, it accelerated only slightly (its speed increased only 13% in the first 10 time steps) while turning away (with 0.034º per time step). In addition, this appeared to be a robust finding, because fast switching oscillating agents turned up in five out of the seven replications.Although the behavioral consequences of the oscillations were evident, the underlying mechanism demanded further analysis. Moreover, the networks from the replicated simulations exhibited oscillations at different frequencies. If the oscillations are adaptive because they enhance the ability to quickly switch between competing behaviors, then the evolutionary demands imposed by the two types of stimuli might induce different oscillation frequencies. After al...

Affective Monitoring: A Generic Mechanism for Affect Elicitation

“…The oscillations facilitated attentional switching from plants to predators, and thus led to quicker and more forceful avoidance reactions when a predator suddenly showed up. In addition, high-frequency oscillations occurred with fitness-increasing stimuli (i.e., food), whereas low-frequency oscillations were evoked by fitness-reducing stimuli (i.e., predators; Heerebout and Phaf, 2010b). …”

“…Competitive models have been applied successfully to self-organization of visual representations (e.g., von der Malsburg, 1973), implicit and explicit memory performance (Murre et al, 1992; Phaf, 1994), attentional selection (Phaf et al, 1990), and even fear conditioning (Armony et al, 1995). A role for competitive processes in the elicitation of affect has first been suggested by the evolutionary simulations of Heerebout and Phaf (2010b). The fact that neural competition emerges so readily when optimizing evolutionary fitness suggests that it may be a basic building block of the neural networks responsible for many kinds of information processing, including cognitive and emotional functioning.…”

“…The networks of the evolved agents revealed strong inhibitory influences between neighboring nodes, which is characteristic of competitive networks (Heerebout and Phaf, 2010b). Lateral inhibition and competition have traditionally served as an explanatory mechanism for selective attention (e.g., Duncan, 1996; see also Phaf et al, 1990), and these simulations indicate that oscillations act as a supplementary mechanism to selection.…”

“…The latter generally imply that priority should be given to steering behavior in more adaptive directions (cf. the evolutionary simulations by Heerebout and Phaf, 2010a,b), which may be avoided if the conflict can be relieved rapidly.…”

Approach, avoidance, and affect: a meta-analysis of approach-avoidance tendencies in manual reaction time tasks

“…Tendencies for appetitive and aversive behaviors in response to positive and negative stimuli, respectively, would thus enhance the adaptation of the organism to its environment. Evolutionary computer simulations have indeed shown that approach and avoidance tendencies toward and away from affective stimuli may emerge autonomously in an organism after a number of generations when starting from a completely random organization (for a more detailed description, see den Dulk et al, 2003; Heerebout and Phaf, 2010a,b). In our daily lives, we are often faced with situations that call for quick and appropriate action.…”