Simple central pattern generator model using phasic analog neurons.

Many biological neurons (called phasic or adapting neurons) display neural adaptation: their response to a constant input diminishes with time. A simple method of adding adaptive firing thresholds to existing analog (or graded-response) neural models is described. A half-center central pattern generator is modeled using two mutually inhibitory phasic analog neurons. Hopf bifurcation analysis shows that oscillatory solutions will arise if the mutual inhibition is sufficiently strong, and allows us to characterize the stability of the cycles which arise.

Different aggressive behaviours are exaggerated by facing vs. broadside subliminal stimuli shown to socially isolated Siamese fighting fish, Betta splendens.

We report and analyse some features of a new phenomenon: socially isolated Betta splendens become extremely hyper-aggressive after seeing brief glimpses of fish models or mirrors. These brief glimpses are below the threshold for releasing aggressive display, so they are considered subliminal aggressive stimuli. The hyper-aggressiveness was observed to last for weeks. To confirm that hyper-aggressiveness was dependent upon the aggressive significance of the subliminal stimuli, we presented socially isolated Betta splendens with subliminal models in either a `facing' posture (used mainly in aggressive contexts), or a `broadside' posture (used in many social contexts). The fish shown the aggressive `facing' subliminal stimuli became more aggressive, while those shown `broadside' stimuli performed more generalized advertisement behaviours. The display posture of the model, which may incorporate specific features relevant to aggression, therefore determined how the subliminal aggressive stimuli altered subsequent aggressiveness. This difference was also persistent. Subliminal stimuli may thus be implicated in the hyper-aggressiveness so often reported after social isolation.

Increased aggressiveness after brief social isolation of adult fish: a connectionist model which organizes this literature.

We examine the literature on fishes' aggressive behaviour after social isolation, in the light of a connectionist adaptive control system model for robot motivation and learning. If animals used the model's motivation modules, then social isolation would cause two simultaneous processes to occur-one progressively increasing motivation, the other decreasing display readiness. The readiness decrement would have the temporal flexibility typical of motivation changes, and would disappear soon after social stimuli reappeared. The incremental effect would be synaptic, and would be temporally more stable. Due to this stability difference, brief post-isolation aggression tests would tend to show that social isolates have depressed attack readiness, but longer tests would uncover the underlying increase in aggressiveness. In reviewing the literature, we find that this has been overwhelmingly the case. A new hypothesis about the adaptive value of the increase of aggression during isolation is outlined, which may help make the phenomenon more understandable.

Social isolation increases social display after priming in Betta splendens but decreases aggressive readiness.

We socially isolated adult, male Siamese fighting fish for 0 to 7 weeks but avoided sensory deprivation, and then measured aggressive display as the fish responded to a series of novel models. Isolates displayed much more strongly than non-isolates to the last model of the series, and display intensity became monotonically stronger with longer social isolation. By contrast, display to the first model seen after isolation was weaker the longer the social isolation. A second experiment compared responses of fish after three weeks during which controls displayed to a conspecific for two minutes every two days, while isolates saw a complex visual stimulus which did not release display. Again, isolates had depressed display readiness, but higher display rates once they were primed. This result supports the hypothesis that social isolation has two effects on aggressive display in Betta, causing decreasing readiness to display, but leading to greater display rates once fish are 'primed'. We discuss two models which predict these phenomena, and show that one of them suggests a new understanding of the function of motivational increases when aggression is 'dammed up'. Methodological issues relevant for examining the rodent behavior literature in the light of the new model are discussed.