Correlation between behavioral and neuronal activities of toads Bufo bufo (L.) in response to moving configurational prey stimuli

Behauioural Processes, 4 (1979) 99-l 06 0 Elsevier Scientific Publishing Company,

99

Amsterdam -Printed

in The Netherlands

CORRELATION BETWEEN BEHAVIORAL AND NEURONAL ACTIVITIES OF TOADS BUFO BUFO (L.) IN RESPONSE TO MOVING CONFIGURATIONAL PREY STIMULI

H.-W. BORCHERS

and J.-P. EWERT

New-o-ethology and Biocybernetic (Federal Republic of Germany)

Laboratories,

University

of Kassel,

D-3500

This work was supported by grants of the Deutsche Forschungsgemeinschaft (Accepted

Kassel

Ew 716.

5 January 1979)

ABSTRACT Borchers, H.-W. and Ewert, J.-P., 1979. Correlation between behavioral and neuronal activities of toads Bufo bufo (L.) in response to moving configurational prey stimuli. Behau. Processes. 4: 99-106. Common toads are able to distinguish prey objects from predators and behaviorally irrelevant stimuli by their shape and direction of motion. Using computer programs for correlation analysis, the prey-catching activity in response to different moving configurational stimuli was compared with the activity of neurons recorded at different levels of the visual pathway. Among retinal ganglion cells, the class R2 neurons were found to be most sensitive, to moving configurational stimuli. Among neurons recorded from retinal projection fields in the optic tectum and thalamic pretectal region, the tectal T5( 2) neurons exhibited configurational selectivity. The output of these neurons showed the best positive correlation with prey-catching when both the neuronal and behavioral activities were compared in response to stripes of different length moving with their axis in, or perpendicular to, the direction of motion.

INTRODUCTION

Animals detect objects of the visual environment and may then respond on the basis of what they see. The decision to respond can be mediated by a releasing mechanism (Lorenz, 1935, 1954), which may be innate (Schleidt, 1962). One of the main functions of sensory recognition systems, such as releasing mechanisms, is the classification of stimulus distributions from the environment into innate and learned classes of functional significance. What chain of neurophysiological events connects a key stimulus with a specific pattern of behavioral responses?

100

Toads Bufo bufo (L.) are able to distinguish prey objects from predators and behaviorally irrelevant stimuli by their shape and direction of motion (Ewert, 1968). Small elongated shapes are treated as prey provided that they move in a direction parallel to their long axis (“worm” configuration). However, the same shapes remain ineffective if they move in a direction perpendicular to their long axis (“antiworm” configuration). The ability of the common toad to distinguish worm-like from antiworm-like moving objects shows a remarkable invariance to changes of other stimulus parameters (Borchers et al., 1978; Ewert et al., 1979). In recent studies the activities of neurons from different levels of the visual pathway in Bufo bufo (L.) have been recorded in response to wormlike and antiworm-like objects (Ewert and Hock, 1972; Ewert and Von Wietersheim, 1974). Using statistical correlation methods, we are interested to know where in the visual system and to what extent the behaviorally important stimulus parameters are encoded. MATERIAL

AND METHODS

The visual stimuli which were presented were rectangular black stripes of constant width (1) and variable length (xl, ) or (x/~) according to whether their axis was oriented parallel or perpendicular to the horizontal direction of movement, with 1= l1 = Z2. In each of the stimulus series, the stimulus started as a small square and the side (xl,) (worm-like objects) or the side (xZZ) (antiworm-like objects) was elongated by steps with x = 1, 2, 4, 8 and 10. All experiments were performed with common toads Bufo b. bufo (L.). In behavioral experiments the value Ewas 2.5 mm’ ; in the neurophysiological studies 1 corresponded to a visual angle of 2”*. The stimulus movement velocity was held constant at u = 25 mm-s-’ in the behavioral experiments and at 7.6 degrees-s-’ in the recording studies. The luminance of the white background against which the black stimuli were moved was 40 cd. mm2in all experiments. The average number of prey-catching orienting responses per min towards the stimulus served as a measure for the discriminative value “prey or no prey” (Ewert, 1968). The average number of impulses per s described the activity of neurons belonging to different classes when the stimuli traversed the centers of their excitatory receptive fields (Ewert and Hock, 1972; Ewert and Von Wietersheim, 1974; cf. also Fig. 1 A and B). ’ The absolute dimension Gebauer, 1973).

was chosen

according

to results

on size constancy

(Ewert

and

Z The dimension of degrees of visual angle was chosen according to results on angular constancy of these neurons under the present recording conditions in paralysed animals. It must be assumed that the visual system is capable of transforming visual angular sizes into absolute dimensions (cf. Ewert et al., 1978).

101

A

TH3 NEURON

B

T5121 NEURON

Fig.1. Records of a single TH3 neuron from the thalamic-pretectal region (A) of the toad Bufo bufo, and a single T5( 2) neuron of the optic tectum (B). The centers of the excitatory receptive fields (ERF) were traversed by stimuli of different configurations: 2” x 2” square; 2” X 8” worm; 8” x 2” antiworm. The black stimuli were moved in a horizontal direction on a white background at 7.6” /s. (From Ewert, 1974.)

The data were expressed as curves and compared using the Pearson waveform correlation (cf. also Sachs, 1976) which allows quantitative comparison of two waveforms. Comparisons were made between the stimulus response relationships of prey-catching behavior (data from Ewert, 1972, cf. 1976) and the stimulus response relationships from single neurons recorded extracellularly at different levels of the visual system, namely retinal ganglion cell classes R2, R3, R4 (from Ewert and Hock, 1972), thalamic-pretectal neurons class TH3 (from Ewert, 1971) and neurons of the optic tectum class T5, response types T5(1) and T5(2) (Von Wietersheim and Ewert, 1978). Continuity was approximated within each curve describing the stimulus response relationships. Data were processed by means of a “Nicolet med 80” computer

102

TABLE

I

Waveform correlation coefficient rw a for comparison of responses to stripes moved in a worm-like (w) and antiworm-like (a)’ manner measured for the prey-catching orienting behavior and for the activity of neurons from retina (R), thalamic-pretectal region (TH) and optic tectum (T). (cf. Fig. 2) System

: Prey-catching

R2

R3

R4

TH3

T5(1)

T5(2)

rw,a

:

-0.8

0.6

0.5

0.8

0.6

-0.6

-0.9

which gave directly

the r values of the Pearson waveform

correlation.

RESULTS

Discrimination

analysis

The activities (R) in response to worm-like (w) and antiworm-like (a) stimuli were compared for the behavior as well as for each neuronal class. This procedure gives first indices for answering the question of how similar is the transformation by the corresponding system of the stimulus parameters xl, and xl,: R = fl(xll)

compared

with R = f2(xlZ)

and thus discriminates between moving configurational The correlation coefficients r,,, are shown in Table I. Waveform

stimuli (cf. Fig. 2).

correlation

The stimulus behavior

response

Rb = fb (Xl,) and R, = were compared relationships.

relationships

(Fig.2)

measured

for prey-catching

f;@h)

with the corresponding

neurophysiological

stimulus

response

Rn = fn(Xl1) and Rn = G(xl~) The question of correlation between behavioral and neurophysiological response activities has to be analyzed for both parameters xl, and xlz. Thus, the correlation coefficient was measured pairwise for worm-like stimuli, rw, and for antiworm-like ones, ra; in the case of correlation, a pair {rw ;ra} = (1; 1) should be expected. A first approach to “linking” behavioral and neurophysiological data was obtained by using “picture correlation” (Borchers, unpubl., cf. Ewert et al., 1978). This method evaluates the amount and waveform of stimulus response relationships. In the present investigation the discrete curves representing the

103

ZTI"UL"S

RESPONSE

RELATIONSHIPS CONTINUITY AND

BY

ITP-50

2110-50

l(O. 30.

20. LO’ 0.

40’ 30. 20.

LO’ 0.

90. 30. 20.

10’ 0.

x

I

.

.

2

Li

.

a

x

20

STIPIULUS

EDGE

I,

b?AGNIFICATION

2

Li FACTOR

6

LO

, X

Fig.2. Stimulus response relationships for prey-catching behavior and for classes of neurons recorded extracellularly from different levels of the visual pathway in the common toad Bufo bufo (L.). Computer processed data. The average activities were measured experimentally in response to stripe stimuli of variable length (magnification factor, X) oriented either in (curve indicated by a white dot) or perpendicular to the horizontal direction of movement (cf. other curve). For explanation see text.

data from neurophysiological and behavioral measurements were compared. In this method, waveform and slope (positive, negative) of the stimulus response relationships are evaluated. The correlation coefficient pairs {‘wp *ra } for neurons belonging to different classes are shown in Table II.

104

TABLE

II

Waveform correlation coefficients TV and r, for comparison between behavioral and neurophysiological responses to worm-like (w) and antiworm-like (a) moving stripes (cf. Fig. 2) Neuronal class Correlation coefficient pair { rw

;

0.5 0.2 0.6 0.9 0.8 0.7

R2 R3 R4 TH3 T5(1) T5(2)

J-J 0.6 -0.7 -0.9 -0.9 0.0 0.9

DISCUSSION

The present investigation confirms previous findings (Ewert et al., 1978) and gives further detailed information on the question of neuronal processing of behaviorally important stimulus parameters.

Retinal ganglion cells The first steps of processing information concerning the shape and movement direction of a stimulus are already performed at the retinal level. Among different classes of retinal ganglion cells investigated, the R2 neurons showed the best selective sensitivity in response to worm-like and antiworm-like moving objects (Table I); also, a relatively high positive correlation coefficient pair was obtained in this class (Table II). It is interesting to note that class R3 and R4 neurons showed even negative correlation coefficients if responses were compared for antiworm-like stimuli (Table II).

Thalamic-pretectal

neurons

Different neuron types have been recorded from the thalamic-pretectal region of the common toad, and many of them are activated mainly by those stimuli which release behavior “avoidance movements” (Ewert, 1971). The class TH3 neurons investigated here (Fig. 1A) exhibit sensitivity in response to configurational stimuli; no correlation was found with the prey-catching activity (c.f. Table II). It should be emphasized that the correlation coefficients for antiworm-like stimuli show even negative values (Table II).

Neurons

of the optic tectum

A variety

of neuron types could also be distinguished

in the optic tectum

105

of the common toad (Ewert and Borchers, 1971; Von Wietersheim and Ewert, 1978). Among class T5 neurons, the response type T5(1) showed sensitivity to moving configurational stimuli (Table I); no correlation was found with the prey-catching activity. The type T5(2), neurons, however, exhibited selective response to configurational stimuli (c.f. Fig. 1B and Table I). Furthermore, relatively high values of {rW ;r,} for positive correlation with the prey-catching activity were found (Table II).

Conclusion The antiworm-configuration is most decisive for determing the response characteristic of a neuron in the present context (Table II). Retinal ganglion cells (class R2) were found to be sensitive to moving configurational stimuli. In projection fields beyond the retinal level, neuronal populations exist showing different kinds of sensitivity in response to worm-like and antiworm-like stimuli (class TH3 and class T5(1)). The selective response characteristic of type T5(2) neurons could be explained by subtractive interactions of T5(1) and TH3 neurons (Ewert, 1974; Ewert and Von Seelen, 1974). It must be emphasized that T5( 2) neurons show no worm-specificity; neither was this seen in the prey-catching behavior. The response spectrum of T5(2) neurons “reflects” approximately the probability that the configuration of a moving stimulus fits the category “prey”, However, the estimation of the absolute size of a stimulus must be achieved by further evaluation processes. We assume that tectal neurons with T5(2) characteristics participate in a system recognizing prey and “commanding” the orienting turn. If a moving stimulus has no prey features, the activity of T5(2) neurons, as well as the probability that the toad orients to it, is relatively low, providing no other components influence the toad’s motivation. For instance: (1) learning processes can modify the innate releasing mechanism (Brzoska and Schneider, 1978); (2) during the mating season the female releases orienting responses by the male (H. Heusser, unpublished, 1966); (3) particular patterns of orienting responses can also be released by moving surroundings, but the opto-kinetic reactions are known to differ from prey-catching orienting movements as regards the key-stimuli as well as the dynamics of response patterns. During future experiments it will be important to investigate whether “gestaltsensitive” neurons, mentioned here, have also to be considered for detection of visual stimuli other than prey. REFERENCES Borchers, H.-W., Burghagen, H. and Ewert, J.-P., 1978. Key stimuli of prey for toads: configuration and movement pattern. J. Comp. Physiol., 128: 189-192. Brzoska, J. and Schneider, H., 1978. Modification of prey-catching behavior by learning in the common toad (Bufo 5. bufo (L.), Anura, Amphibia): changes in responses to visual objects and effects of auditory stimuli. Behav. Processes, 3: 125-136. Ewert, J.P., 1968. Der Einfluss von Zwischenhirndefekten auf die Visuomotorik im Beute- und Fluchtverhalten der ErdkrGte (Bufo bufo L.). Z. Vergl. Physiol., 61: 41-70.

106 Ewert, J.-P., 1971. Single unit response of the toad’s (Bufo americanus) caudal thalamus to visual objects. Z. Vergl. Physiol., 74: 81-102. Ewert, J.-P., 1972. Zentralnervose Analyse und Verarbeitung visueller Sinnesreize. Naturwiss. Rundsch., 25: l-11. Ewert, J.-P., 1974. The neural basis of visually guided behavior. Sci. Am., 230: 34-42. Ewert, J.-P., 1976. The visual system of the toad: behavioral and physiological studies on a pattern recognition system. In: K.V. Fite (Editor), The Amphibian Visual System: a Multidisciplinary Approach. Academic Press, New York, N.Y., San Francisco, Calif., London, pp. 141-202. Ewert, J.-P., and Borchers, H.-W., 1971. Reaktionscharakteristik von Neuronen aus dem Tectum opticum und Subtectum der ErdkrGte Bufo bufo (L.). Z. Vergl. Physiol., 71: 165-189. Ewert, J.-P. and Gebauer, L., 1973. Griissenkonstanzphiinomene im Beutefangverhalten der ErdkrSte (Bufo bufo L.). Z. Vergl. Physiol., 85: 303-315. Ewert, J.-P,, and Hock, F.J., 1972. Movement sensitive neurones in the toad’s retina. Exp. Brain. Res., 16: 41-59. Ewert, J.-P. and Von Seelen, W., 1974. Neurobiologie und System-Theorie eines visuellen Mustererkennungsmechanismus bei K&en. Kybernetik, 14: 167-183. Ewert, J.-P. and Von Wietersheim, A., 1974. Musterauswertung durch Tectum- und Thalamus/Prlitectum-Neurone im visuellen System der KrSte Bufo bufo (L.). J. Comp. Physiol., 92: 131-148. Ewert, J.-P., Borchers, H.-W. and Von Wietersheim, A., 1978. Question of prey feature detectors in the toad’s Bufo bufo (L.) visual system: a correlation analysis. J. Comp. Physiol., 126: 43-47. Ewert, J.-P., Arend, B., Becker, V. and Borchers, H.-W., 1979. Invariants in configurational prey-selection by Bufo bufo (L.). Brain Behav. Evol. (in press). Lorenz, K., 1935. Der Kumpan in der Umwelt des Vogels. J. Ornithol., 83: 137-413. Lorenz, K., 1954. Das angeborene Erkennen. Natur und Volk, 84: 285-295. Sachs, L., 1976. Statist&he Methoden. Springer, Berlin, Heidelberg, New York, N.Y., 105 pp. Schleidt, W., 1962. Die historische Entwicklung der Begriffe “Angeborenes auslosendes Schema” und “Angeborener AuslSsemechanismus” in der Ethologie. Z. Tierpsychol., 19: 697-722. Von Wietersheim, A. and Ewert, J.-P., 1978. Neurons of the toad’s (Bufo bufo L.) visual system sensitive to moving configurational stimuli: a statistical analysis. J. Comp. Physiol., 126: 35-42.