How electroreceptors encode JAR-eliciting stimulus regimes: Reading trajectories in a phase-amplitude plane

Abstract

This study correlates certain features of the Jamming Avoidance Response (JAR) in the electric fish,Eigenmannia, with specific properties of electroreceptors and higher-order electrosensitive units. The JAR is a shift of the animal's own electric organ discharge (EOD) frequency away from similar frequencies of foreign signals, such as EODs of conspecifics. It is controlled by characteristic modulations in the amplitude, ¦S¦, and phase,H, of the electric stimulus which result if foreign signals are superimposed upon the animal's own EOD. Modulations of ¦S¦ andH are respectively encoded in terms of the probability,p, ofP-type electroreceptor firing and in the timing,t, ofT-type electroreceptor firing (Fig. 1). In this paper we describe experiments on curarized animals whose silenced EODs were replaced by an electric stimulus,S1, while other stimuli,S2 and sometimes an additionalS3, mimicked EODs of conspecific neighbors. The results are as follows: 1. The complicated modulations in phase,H, and amplitude, ¦S¦, which result if two foreign stimuli interfere with the animal's EOD, are rather faithfully represented by the timing,t, ofT-unit firing and the probability,p, ofP-unit firing respectively (Fig. 2). However, larger systematic distortions are found in the representation of certain stimulus regimes (Fig. 3), and these distortions cause specific behavioral effects. 2. Given an EOD fundamental frequency,f, JARs can be elicited by a sine wave stimulus with frequency near that of either the fundamental or the higher harmonic, 2f. No JARs result if the stimulus frequency is near the subharmonic,f/2 (Bullock et al. 1972). Correspondingly, the probability,p, ofP-unit firing and the timing,t, ofT-unit firing are modulated in a continuous manner only in the first two instances (f and 2f) but in a discontinuous manner in the last case (f/2). Thus only continuous modulations ofp andt appear to elicit the JAR (Figs. 4 to 6). 3. The possibility of viewing the effect of a closed graph in the ¦S¦,H-state plane as the sum of the contributions made by small vectors which constitute this graph is discussed in the light of receptor responses (Figs. 7–9). 4. If the intensity ratio ¦S2¦/¦S1¦ is constant over the animal's body surface no JAR can be elicited as long asS1 andS2 are pure sine waves. However, JARs will result under this condition ifS1 is clipped in one half cycle of its sinusoidal wave from. This correlates with the fact that the modulations of the probability,p, ofP-unit firing and the modulation of the difference,t1-t2, in the timing of the firing of pairs ofT-units become Δf-sign specific ifS1 is clipped. This sign specificity introduces sign-specific differences inp, t-graphs as they normally, and in much stronger form, result from local differences in the ¦S2¦/¦S1¦ ratio (Figs. 10, 11). It is therefore not necessary to postulate JAR mechanisms other than the one presented earlier (Heiligenberg et al. 1978) in order to explain the effect of clipped sine waves.