Perception, 1977, volume 6, pages 513-527
Developmental constraints of motion detection mechanisms in the kitten Jean-Marc Flandrin, Marc Jeannerod I.N.S.E.R.M. - Unite 94, Laboratoire de Neuropsychologie Experimental, 16 Avenue du Doyen Lepine, 69500 Bron, France Presented at a symposium on Sensory Psychophysiology of the Visual Perception of Movement, held in parallel with the XXI International Congress of Psychology, Paris, July 1976 (received in revised form 9 March 1977)
Abstract. The influence p'E deprivation procedures on the development of motion detection mechanisms has been studied in twenty-two kittens. Superior colliculus neurons did not acquire direction selectivity and normal ocular dominance in animals reared in the dark or in stroboscopic light. Neuron immaturity persisted in spite of a five week additional recovery period in normal conditions. Exposure to unidirectional visual motion for 10 h during the fifth week of postnatal age produced an asymmetric development of the two superior colliculi. Finally, unilateral neonatal ablation of visual cortex permanently impaired development of the ipsilateral superior colliculus. In the same or in different animals, development of optokinetic nystagmus, a typical visuomotor response, was similarly influenced by the global or selective deprivation procedures. These results suggest that motion detection mechanisms (both afferent and efferent) strongly depend upon constraints imposed by the visual world during the first weeks of life. 1 Introduction Motion detection seems to be subserved by a specific class of neuronal detectors sensitive to motion, that provide information as to the direction and the velocity of moving objects. Efferent mechanisms generate visuomotor responses adapted to the movement detected by the afferent system. Physiological and behavioural aspects of motion detection have been the subject of a number of studies in the past fifteen years. Neurons sensitive to motion, i.e. tuned for a given direction of motion, have been located centrally at the cortical (Hubel and Wiesel 1962) and tectal (Marchiafava and Pepeu 1966) levels. Cortical neurons might subserve a precise analysis of movement patterns within given areas of the visual field. Neurons from the superior colliculus (at least in the cat) would be more specifically devoted to detection of the displacement of objects across the visual field. Unlike cortical neurons they have relatively large receptive fields; they are virtually insensitive to the shape or the orientation of the stimulus; they do not respond to stationary stimuli. In the present study we deal exclusively with those neurons located in the upper and intermediate layers of superior colliculus, and with their afferents from the visual cortex. Motor behaviour related to motion detection is also well documented. Optokinetic nystagmus (OKN), an easily recordable response present in all normal cats, is a typical oculomotor behaviour triggered by the continuous displacement of large visual scenes. The fact that OKN will be studied here in parallel to collicular neurons does not imply a structure-to-function relationship between superior colliculus and OKN. Even though this relationship has been postulated by several authors (see discussion), we only consider in the present context that collicular neuronal responses to motion and OKN, respectively, represent the beginning and the end of the same complex process. We assume here that development of visual functions in the cat depends, at least in part, on adequate stimulation by the environment. Neurons from the visual cortex,
514
J-M Flandrin, M Jeannerod
for instance, fail to acquire orientation specificity in kittens reared in the dark for their first weeks of life (e.g. Wiesel and Hubel 1965a). The same neurons can get monotonically tuned by exposing the young animal to an orientationally biased environment (e.g. Hirsch and Spinelli 1970; Blakemore and Cooper 1970). To what extent these developmental constraints also apply to collicular neurons, and whether development of OKN parallels acquisition of normal properties by the superior colliculi, are our basic questions in this paper. [Preliminary notes have already been published (Flandrin and Jeannerod 1975, 1977; Flandrin et al 1976).] 2 Material and methods 2.1 Rearing conditions A total of twenty-two kittens were used. Animals were born in the laboratory and were directed during the first postnatal week (i.e. before opening of the eyes) to different rearing conditions. Normal conditions (six kittens, control group). Animals were kept with their mother in usual laboratory cages, until the day of testing. Rearing in the dark (seven kittens, dark-reared group). Litters were put in separate cages in a lightproof room. Feeding and cleaning were done in the dark. Four kittens remained in the dark until the day of testing (twelve weeks). The three others underwent a light deprivation period of twenty weeks from birth, and were then returned to a normally lighted environment for five more weeks before being tested (recovery group). Rearing in stroboscopic light (three kittens, strobe group) (1) . Animals maintained in an otherwise dark room were exposed to stroboscopic illumination for 12 h d"1. Flashes frequency was 2 s"1 and flash duration, 0-2 ms. Duration of strobe rearing was ten months from birth. This condition is known to suppress retinal signals of motion (self-produced, or arising from the external world). Rearing in a unidirectionally moving environment (four kittens, unidirectional group). Kittens were reared in total darkness from birth up to the end of the fourth week. During the fifth week each kitten was exposed for 10 h (2 h d"1, 5 d) to a unidirectionally moving visual scene, in the apparatus described by Vital-Durand and Jeannerod (1974a). Velocity of the rotating visual scene was 20° s"1 and direction of rotation was right ward. Kittens were kept in darkness between the exposure sessions and after exposure time had elapsed, until the day of testing. Exposure to rightward unidirectional motion can be understood either as a selective enhacement of retinal rightward-motion signals, or as a selective deprivation of left ward-motion signals. Neonatal visual decortication (five kittens, decorticated group). Animals underwent a unilateral right ablation of posterior visual cortex on the day after birth. Lesion was made by aspiration under ketamine-base anaesthesia (15 mg kg -1 ; intramuscularly). Animals were kept in normal conditions until the day of testing. 2.2 Unit recording Results reported here were drawn from a sample of 431 neurons recorded from the superior colliculus. For recording, animals were prepared under anaesthesia with an initial dose of Nembutal (35 mg kg -1 , intraperitoneally; dark-reared and unidirectional groups), or under ketamine (15 mg kg"1, intramuscularly; control, strobe, decorticated groups). An intravenous catheter and a tracheal cannula were placed. ^ Kittens from the strobe group were kindly provided by B Amblard, Institut de Neurophysiologie et Psychophysiologic (Pr Paillard), CNRS (Marseille).
Motion detection mechanisms in the kitten
515
Animals were positioned in a stereotaxic frame and paralyzed by an initial injection of flaxedil (20 mg kg"1) and artificially ventilated. During the experiment a light anaesthesia was maintained [Nembutal (4 mg kg -1 h" 1 )]. Points of contact and wounds caused by the placement of the cannula and the catheter were frequently injected with 5% xylocaine. A continuous injection of flaxedil [(5 mg kg -1 ) h" 1 ] was given throughout the experiment. The temperature of the animal was maintained between 37°C and 38°C by means of a warming blanket. The lids were maintained retracted by 15% neosynephrine and pupils were dilated by 1% atropine. Corneas were prevented from drying by corneal lenses. The skull was opened' over the superior colliculus area. The cortex was exposed and protected by a film of agar in 0-9% saline. Glass platinum electrodes were lowered vertically within the collicular region, according to the stereotaxic coordinates established by Rose and Goodfellow (1973). Neurons were stimulated by variations of background illumination, and by manual displacement of luminous spots or edges projected onto a screen in front of the kittens' eyes. Dark spots or edges on a white background could also be used. Responses to moving stimuli and to diffuse light, and ocular dominance are considered in these experiments. Responses to moving stimuli were classified into three groups according to the audible response. Direction selectivity refers to a response to a moving stimulus crossing the receptive field, obtained in one direction only. Direction preference refers to a larger response for one direction of the moving stimulus without evidence for a null direction. Nondirectional response refers to neurons responding indifferently to stimuli moving in any direction. Binocularity was estimated subjectively. Alternative eye stimulation allowed the neurons to be classified in one of five groups. Group 1 contains the cells driven only by the contralateral eye; group 3, the cells driven equally by the two eyes, and group 5, the cells driven only by the ipsilateral eye. The other groups (2 and 4) are intermediate. 2.3 Optokinetic nystagmus OKN was recorded in the same animals as those used for unit recordings. Under Nembutal anaesthesia (35 mg kg"1, intraperitoneally), electrodes for recording eye movements, and a head fixation device were fixed to the skull. Recording sessions started a few days after surgery. They were made as short as possible (as a rule, less than 15 min for one kitten) so as to avoid interference with the rearing conditions, and to prevent a possible functional restoration before study of the superior colliculus. The animal was placed within a cylindric chamber (1 m diameter). The head was secured to a hammock, and positioned at the centre of the cylinder. Horizontal eye movements were recorded with a DC amplifier (0-50 Hz). Stimulation was provided by rotating clockwise (CW) or counterclockwise (CCW) the wall of the cylinder, which was equipped with black and white vertical stripes (10 deg of visual angle per stripe). Speed of rotation could be varied. Results were expressed as the frequency of beats per 25 s epoch, versus angular velocity of the wall of the cylinder. 3 Results 3.1 Motion detection in the ten-twenty-week-old normal kitten Superior colliculus (87 neurons). In kittens from the 'control group', neurons had little or no spontaneous activity. Changes in background illumination elicited brief responses in most cases of the ON type. Figure la shows the repartition of the different types of responses to moving stimuli in 87 neurons. Of these 73 (84%) displayed a directional response, with direction selectivity in 39 cases (45%) and direction preference in 34 cases (39%). In most cases, preferred direction was away
516
J-M Flandrin, M Jeannerod
from the fovea (e.g. centrifugal in the visual field), resulting in the classical predominant leftward response of the right colliculus, and rightward response of the left colliculus (Straschill and Hoffman 1968). Finally, in 14 cases (16%) responses to moving stimuli were nondirectional. When compared with the results of other authors, the number of directional responses appears higher in our experiments. Berman and Cynader (1972) and Hoffmann (1973), in adult cats, considered only 65% and 57% of responses in their respective samples to be directional. We think that this difference is a matter of definition. The authors cited above used a quantitative estimation of the response and classified as 'directional' neurons which give in their preferred direction a response more than twice that in the opposite direction. In our studies we used a different strategy. Since we feel that the distinction between directional and nondirectional responses represents a qualitative rather than a quantitative difference, we classified separately neurons with a directional bias (according to the intensity of the response in the preferred direction) and neurons without directional bias. Results of the ocular dominance tests are given in figure lb. Of the 54 neurons tested, 21 (39%) were driven exclusively or more efficiently by the contralateral eye, 25 (46%) were driven equally by both eyes, and 8 (15%) by the ipsilateral eye. These values are in accordance with those of Berman and Cynader (1972) and Hoffmann (1973). 80
N=S1
® 60 g 40 20
1 2
3
1 2 3 4 5 Contralateral *- equal -> ipsilateral
(a) (b) Figure 1. Directional responses and ocular dominance in superior colliculus neurons from normal four-month-old kittens, (a) Types of response to movement (%): 1, direction-specific responses; 2, direction preference; 3, nondirectional responses, (b) Degree of ocular dominance (%): 1, response to stimulation of the contralateral eye only; 3, equivalent response of both eyes; 5, response to stimulation of the ipsilateral eye only; 2 and 4 are intermediate.
50
50-
£ 40 o
40-
A---AN->T
o-oT->N a
>» g 30
30-
P
20-
£ 20 •AV-
20 Velocity f s " 1 ) binocular
30
u
20 Velocity (° s"1) left eye
30
20 Velocity (° s_1) right eye
Figure 2. Optokinetic responses in a four-month-old normal kitten. Frequency of nystagmic beats per time units of 25 s are plotted against the angular velocity of the drum, under binocular, left-eye, and right-eye stimulation. N ->• T: drum rotates from nasal to temporal, i.e. clockwise for the right eye, counterclockwise for the left eye; T -> N: drum rotates from temporal to nasal.
517
Motion detection mechanisms in the kitten
Optokinetic nystagmus. In normal kittens, recorded with binocular viewing, OKN frequency and stimulus velocity could be positively correlated up to 50° s"1. When one eye was occluded, OKN frequency was higher for a displacement of the stripes in the temporal-to-nasal direction (e.g. clockwise for the right eye) (figure 2). This effect, although much clearly observed in rabbits or guinea pigs, has also been noted in the cat by Braun and Gault (1969), and by Wood et al (1973). 3.2 Influence of dark rearing on the development of motion detection mechanisms Superior colliculus (143 neurons:42 neurons from the 'dark-reared group', 101 neurons from the 'recovery group'). In kittens from the 'dark-reared group', raised without visual experience, the response pattern of the 42 neurons recorded from the superior colliculus was quite different from that of normals. There was more spontaneous activity. Changes in background illumination elicited strong responses in all cases, mostly of the ON-OFF type. Almost all neurons responded to moving spots displaced at low velocities. This response showed a directional bias in 7 cases only out of 42 (17%). In all the remaining cases (35 out of 42, or 83%) responses were found to be nondirectional according to our criteria (figure 3a). As this group was tested at an early stage of the study, data on binocularity are not available. However, in a similar experiment in kittens whose lids were sutured bilaterally during the first postnatal week, and recorded ten months later, Hoffmann and Sherman (1975) found 60% of visually responsive superior colliculus neurons driven exclusively or more efficiently by the contralateral eye, 22% equally by both eyes, and only 9% by the ipsilateral eye (figure 3b). In three other kittens ('recovery group') 101 neurons were recorded from the superior colliculus after an additional five week period spent in normal laboratory conditions. Neuronal responses were found to be closely similar to those from the 'dark-reared group', recorded immediately at the end of the dark period. Of the 101 neurons studied in the three kittens, 32 (31%) had no response to flashed diffuse light. The 69 (69%) remaining had a weak response, in most cases of the ON type (59 out of 69). The results concerning the response to moving stimuli are shown in figure 4a: 59 neurons (59%) were classified as nondirectional; 30 (30%) had a direction preference, and only 12 neurons with a direction selective response were found. Results of the ocular dominance test are illustrated in figure 4b, which shows that 46 neurons (46%) were monocular and driven only by the contralateral eye, 33 (33%) were binocular with a greater contribution from the contralateral eye, 21 (21%) were binocular and equally driven by both eyes. Only one neuron was found to be binocular with a preference for the ipsilateral eye. # = 42
DU
|
^m
7V=118
1 2 3 4 5 Contralateral «- equal -+ ipsilateral
(a) (b) Figure 3. Directional responses and ocular dominance in superior colliculus neurons from darkreared kittens, (a) Distribution of 42 neurons from the 'dark-reared group', (b) Redrawn from the results of Hoffmann and Sherman (1975). Distribution of 118 neurons recorded in ten-monthold kittens lid-sutured at birth. Legend as in figure 1.
518
J-M Flandrin, M Jeannerod
Optokinetic nystagmus. As already mentioned by Vital-Durand et al (1974) OKN is present in visually naive ten-fifteen-week-old kittens, when they are first put in the testing situation. This result, confirmed by the present study, is controversial to that of Van Hof-Van Duin (1976). This author, using a visual inspection technique to assess eye movements, found that five to nine days spent in a normally lighted environment were necessary for OKN to appear in four-month-old kittens dark-reared from birth. In our dark-reared animals, however, OKN was far from normal. Saccades were large, and somewhat erratically distributed in time. The main difference with respect to normal controls was that beat frequency was not positively correlated with stimulus velocity. Instead, the frequency of beats steeply decreased when velocity was increased (figure 5). In two kittens from the 'recovery group', OKN was studied up to the third week after the end of the dark period. No improvement was observed.
W=101
#=101
l
2
1 2 3 4 5 Contralateral *- equal -> ipsilateral
3
(a) (b) Figure 4. Directional responses and ocular dominance in superior colliculus neurons after five week 'recovery' period from dark rearing. Same legend as in figure 1.
40
a*
2 30
* -\V
20
Velocity (° s"1)
50
Figure 5. Optokinetic responses in two dark-reared kittens, recorded in the binocular condition. Same legend as in figure 2.
3.3 Influence of rearing in stroboscopic light Superior colliculus (110 neurons). In the 110 neurons studied from the 'strobe group', the spontaneous activity was low. Eighty-nine neurons (81%) had no response to flashed diffuse light; the remaining 21 (19%) had a weak response. The results of the directionality tests are shown in figure 6a. Only two direction-selective neurons
Motion detection mechanisms in the kitten
519
were found; 11 (10%) had a direction preference, and the majority (97, or 88%) were nondirectional. The results of the ocular dominance test, illustrated in figure 6b, show that 68 neurons (62%) were monocular and driven only by the contralateral eye; 14 (13%) were binocular, with a greater contribution from the contralateral eye; and 27 (24%) were binocular and equally driven by both eyes. Only one neuron was found to be binocular, with a greater tendency to be driven by the ipsilateral eye. These results show that the collicular neurons of strobe-reared cats, i.e. with no experience of visual motion, have directionality and binocularity characteristics similar to those of dark-reared and binocularly sutured cats. Optokinetic nystagmus. OKN was even weaker in the 'strobe group' than in the 'dark-reared group', in the sense that it was less readily obtained when the animals were first placed in the testing situation. OKN displayed the same characteristic as in the 'dark-reared group', i.e., the frequency of beats decreased when the velocity of the stimulus increased. 7V=110
JV=110
l
2
3
1 2 3 4 5 Contralateral
Developmental constraints of motion detection mechanisms in the kitten Jean-Marc Flandrin, Marc Jeannerod I.N.S.E.R.M. - Unite 94, Laboratoire de Neuropsychologie Experimental, 16 Avenue du Doyen Lepine, 69500 Bron, France Presented at a symposium on Sensory Psychophysiology of the Visual Perception of Movement, held in parallel with the XXI International Congress of Psychology, Paris, July 1976 (received in revised form 9 March 1977)
Abstract. The influence p'E deprivation procedures on the development of motion detection mechanisms has been studied in twenty-two kittens. Superior colliculus neurons did not acquire direction selectivity and normal ocular dominance in animals reared in the dark or in stroboscopic light. Neuron immaturity persisted in spite of a five week additional recovery period in normal conditions. Exposure to unidirectional visual motion for 10 h during the fifth week of postnatal age produced an asymmetric development of the two superior colliculi. Finally, unilateral neonatal ablation of visual cortex permanently impaired development of the ipsilateral superior colliculus. In the same or in different animals, development of optokinetic nystagmus, a typical visuomotor response, was similarly influenced by the global or selective deprivation procedures. These results suggest that motion detection mechanisms (both afferent and efferent) strongly depend upon constraints imposed by the visual world during the first weeks of life. 1 Introduction Motion detection seems to be subserved by a specific class of neuronal detectors sensitive to motion, that provide information as to the direction and the velocity of moving objects. Efferent mechanisms generate visuomotor responses adapted to the movement detected by the afferent system. Physiological and behavioural aspects of motion detection have been the subject of a number of studies in the past fifteen years. Neurons sensitive to motion, i.e. tuned for a given direction of motion, have been located centrally at the cortical (Hubel and Wiesel 1962) and tectal (Marchiafava and Pepeu 1966) levels. Cortical neurons might subserve a precise analysis of movement patterns within given areas of the visual field. Neurons from the superior colliculus (at least in the cat) would be more specifically devoted to detection of the displacement of objects across the visual field. Unlike cortical neurons they have relatively large receptive fields; they are virtually insensitive to the shape or the orientation of the stimulus; they do not respond to stationary stimuli. In the present study we deal exclusively with those neurons located in the upper and intermediate layers of superior colliculus, and with their afferents from the visual cortex. Motor behaviour related to motion detection is also well documented. Optokinetic nystagmus (OKN), an easily recordable response present in all normal cats, is a typical oculomotor behaviour triggered by the continuous displacement of large visual scenes. The fact that OKN will be studied here in parallel to collicular neurons does not imply a structure-to-function relationship between superior colliculus and OKN. Even though this relationship has been postulated by several authors (see discussion), we only consider in the present context that collicular neuronal responses to motion and OKN, respectively, represent the beginning and the end of the same complex process. We assume here that development of visual functions in the cat depends, at least in part, on adequate stimulation by the environment. Neurons from the visual cortex,
514
J-M Flandrin, M Jeannerod
for instance, fail to acquire orientation specificity in kittens reared in the dark for their first weeks of life (e.g. Wiesel and Hubel 1965a). The same neurons can get monotonically tuned by exposing the young animal to an orientationally biased environment (e.g. Hirsch and Spinelli 1970; Blakemore and Cooper 1970). To what extent these developmental constraints also apply to collicular neurons, and whether development of OKN parallels acquisition of normal properties by the superior colliculi, are our basic questions in this paper. [Preliminary notes have already been published (Flandrin and Jeannerod 1975, 1977; Flandrin et al 1976).] 2 Material and methods 2.1 Rearing conditions A total of twenty-two kittens were used. Animals were born in the laboratory and were directed during the first postnatal week (i.e. before opening of the eyes) to different rearing conditions. Normal conditions (six kittens, control group). Animals were kept with their mother in usual laboratory cages, until the day of testing. Rearing in the dark (seven kittens, dark-reared group). Litters were put in separate cages in a lightproof room. Feeding and cleaning were done in the dark. Four kittens remained in the dark until the day of testing (twelve weeks). The three others underwent a light deprivation period of twenty weeks from birth, and were then returned to a normally lighted environment for five more weeks before being tested (recovery group). Rearing in stroboscopic light (three kittens, strobe group) (1) . Animals maintained in an otherwise dark room were exposed to stroboscopic illumination for 12 h d"1. Flashes frequency was 2 s"1 and flash duration, 0-2 ms. Duration of strobe rearing was ten months from birth. This condition is known to suppress retinal signals of motion (self-produced, or arising from the external world). Rearing in a unidirectionally moving environment (four kittens, unidirectional group). Kittens were reared in total darkness from birth up to the end of the fourth week. During the fifth week each kitten was exposed for 10 h (2 h d"1, 5 d) to a unidirectionally moving visual scene, in the apparatus described by Vital-Durand and Jeannerod (1974a). Velocity of the rotating visual scene was 20° s"1 and direction of rotation was right ward. Kittens were kept in darkness between the exposure sessions and after exposure time had elapsed, until the day of testing. Exposure to rightward unidirectional motion can be understood either as a selective enhacement of retinal rightward-motion signals, or as a selective deprivation of left ward-motion signals. Neonatal visual decortication (five kittens, decorticated group). Animals underwent a unilateral right ablation of posterior visual cortex on the day after birth. Lesion was made by aspiration under ketamine-base anaesthesia (15 mg kg -1 ; intramuscularly). Animals were kept in normal conditions until the day of testing. 2.2 Unit recording Results reported here were drawn from a sample of 431 neurons recorded from the superior colliculus. For recording, animals were prepared under anaesthesia with an initial dose of Nembutal (35 mg kg -1 , intraperitoneally; dark-reared and unidirectional groups), or under ketamine (15 mg kg"1, intramuscularly; control, strobe, decorticated groups). An intravenous catheter and a tracheal cannula were placed. ^ Kittens from the strobe group were kindly provided by B Amblard, Institut de Neurophysiologie et Psychophysiologic (Pr Paillard), CNRS (Marseille).
Motion detection mechanisms in the kitten
515
Animals were positioned in a stereotaxic frame and paralyzed by an initial injection of flaxedil (20 mg kg"1) and artificially ventilated. During the experiment a light anaesthesia was maintained [Nembutal (4 mg kg -1 h" 1 )]. Points of contact and wounds caused by the placement of the cannula and the catheter were frequently injected with 5% xylocaine. A continuous injection of flaxedil [(5 mg kg -1 ) h" 1 ] was given throughout the experiment. The temperature of the animal was maintained between 37°C and 38°C by means of a warming blanket. The lids were maintained retracted by 15% neosynephrine and pupils were dilated by 1% atropine. Corneas were prevented from drying by corneal lenses. The skull was opened' over the superior colliculus area. The cortex was exposed and protected by a film of agar in 0-9% saline. Glass platinum electrodes were lowered vertically within the collicular region, according to the stereotaxic coordinates established by Rose and Goodfellow (1973). Neurons were stimulated by variations of background illumination, and by manual displacement of luminous spots or edges projected onto a screen in front of the kittens' eyes. Dark spots or edges on a white background could also be used. Responses to moving stimuli and to diffuse light, and ocular dominance are considered in these experiments. Responses to moving stimuli were classified into three groups according to the audible response. Direction selectivity refers to a response to a moving stimulus crossing the receptive field, obtained in one direction only. Direction preference refers to a larger response for one direction of the moving stimulus without evidence for a null direction. Nondirectional response refers to neurons responding indifferently to stimuli moving in any direction. Binocularity was estimated subjectively. Alternative eye stimulation allowed the neurons to be classified in one of five groups. Group 1 contains the cells driven only by the contralateral eye; group 3, the cells driven equally by the two eyes, and group 5, the cells driven only by the ipsilateral eye. The other groups (2 and 4) are intermediate. 2.3 Optokinetic nystagmus OKN was recorded in the same animals as those used for unit recordings. Under Nembutal anaesthesia (35 mg kg"1, intraperitoneally), electrodes for recording eye movements, and a head fixation device were fixed to the skull. Recording sessions started a few days after surgery. They were made as short as possible (as a rule, less than 15 min for one kitten) so as to avoid interference with the rearing conditions, and to prevent a possible functional restoration before study of the superior colliculus. The animal was placed within a cylindric chamber (1 m diameter). The head was secured to a hammock, and positioned at the centre of the cylinder. Horizontal eye movements were recorded with a DC amplifier (0-50 Hz). Stimulation was provided by rotating clockwise (CW) or counterclockwise (CCW) the wall of the cylinder, which was equipped with black and white vertical stripes (10 deg of visual angle per stripe). Speed of rotation could be varied. Results were expressed as the frequency of beats per 25 s epoch, versus angular velocity of the wall of the cylinder. 3 Results 3.1 Motion detection in the ten-twenty-week-old normal kitten Superior colliculus (87 neurons). In kittens from the 'control group', neurons had little or no spontaneous activity. Changes in background illumination elicited brief responses in most cases of the ON type. Figure la shows the repartition of the different types of responses to moving stimuli in 87 neurons. Of these 73 (84%) displayed a directional response, with direction selectivity in 39 cases (45%) and direction preference in 34 cases (39%). In most cases, preferred direction was away
516
J-M Flandrin, M Jeannerod
from the fovea (e.g. centrifugal in the visual field), resulting in the classical predominant leftward response of the right colliculus, and rightward response of the left colliculus (Straschill and Hoffman 1968). Finally, in 14 cases (16%) responses to moving stimuli were nondirectional. When compared with the results of other authors, the number of directional responses appears higher in our experiments. Berman and Cynader (1972) and Hoffmann (1973), in adult cats, considered only 65% and 57% of responses in their respective samples to be directional. We think that this difference is a matter of definition. The authors cited above used a quantitative estimation of the response and classified as 'directional' neurons which give in their preferred direction a response more than twice that in the opposite direction. In our studies we used a different strategy. Since we feel that the distinction between directional and nondirectional responses represents a qualitative rather than a quantitative difference, we classified separately neurons with a directional bias (according to the intensity of the response in the preferred direction) and neurons without directional bias. Results of the ocular dominance tests are given in figure lb. Of the 54 neurons tested, 21 (39%) were driven exclusively or more efficiently by the contralateral eye, 25 (46%) were driven equally by both eyes, and 8 (15%) by the ipsilateral eye. These values are in accordance with those of Berman and Cynader (1972) and Hoffmann (1973). 80
N=S1
® 60 g 40 20
1 2
3
1 2 3 4 5 Contralateral *- equal -> ipsilateral
(a) (b) Figure 1. Directional responses and ocular dominance in superior colliculus neurons from normal four-month-old kittens, (a) Types of response to movement (%): 1, direction-specific responses; 2, direction preference; 3, nondirectional responses, (b) Degree of ocular dominance (%): 1, response to stimulation of the contralateral eye only; 3, equivalent response of both eyes; 5, response to stimulation of the ipsilateral eye only; 2 and 4 are intermediate.
50
50-
£ 40 o
40-
A---AN->T
o-oT->N a
>» g 30
30-
P
20-
£ 20 •AV-
20 Velocity f s " 1 ) binocular
30
u
20 Velocity (° s"1) left eye
30
20 Velocity (° s_1) right eye
Figure 2. Optokinetic responses in a four-month-old normal kitten. Frequency of nystagmic beats per time units of 25 s are plotted against the angular velocity of the drum, under binocular, left-eye, and right-eye stimulation. N ->• T: drum rotates from nasal to temporal, i.e. clockwise for the right eye, counterclockwise for the left eye; T -> N: drum rotates from temporal to nasal.
517
Motion detection mechanisms in the kitten
Optokinetic nystagmus. In normal kittens, recorded with binocular viewing, OKN frequency and stimulus velocity could be positively correlated up to 50° s"1. When one eye was occluded, OKN frequency was higher for a displacement of the stripes in the temporal-to-nasal direction (e.g. clockwise for the right eye) (figure 2). This effect, although much clearly observed in rabbits or guinea pigs, has also been noted in the cat by Braun and Gault (1969), and by Wood et al (1973). 3.2 Influence of dark rearing on the development of motion detection mechanisms Superior colliculus (143 neurons:42 neurons from the 'dark-reared group', 101 neurons from the 'recovery group'). In kittens from the 'dark-reared group', raised without visual experience, the response pattern of the 42 neurons recorded from the superior colliculus was quite different from that of normals. There was more spontaneous activity. Changes in background illumination elicited strong responses in all cases, mostly of the ON-OFF type. Almost all neurons responded to moving spots displaced at low velocities. This response showed a directional bias in 7 cases only out of 42 (17%). In all the remaining cases (35 out of 42, or 83%) responses were found to be nondirectional according to our criteria (figure 3a). As this group was tested at an early stage of the study, data on binocularity are not available. However, in a similar experiment in kittens whose lids were sutured bilaterally during the first postnatal week, and recorded ten months later, Hoffmann and Sherman (1975) found 60% of visually responsive superior colliculus neurons driven exclusively or more efficiently by the contralateral eye, 22% equally by both eyes, and only 9% by the ipsilateral eye (figure 3b). In three other kittens ('recovery group') 101 neurons were recorded from the superior colliculus after an additional five week period spent in normal laboratory conditions. Neuronal responses were found to be closely similar to those from the 'dark-reared group', recorded immediately at the end of the dark period. Of the 101 neurons studied in the three kittens, 32 (31%) had no response to flashed diffuse light. The 69 (69%) remaining had a weak response, in most cases of the ON type (59 out of 69). The results concerning the response to moving stimuli are shown in figure 4a: 59 neurons (59%) were classified as nondirectional; 30 (30%) had a direction preference, and only 12 neurons with a direction selective response were found. Results of the ocular dominance test are illustrated in figure 4b, which shows that 46 neurons (46%) were monocular and driven only by the contralateral eye, 33 (33%) were binocular with a greater contribution from the contralateral eye, 21 (21%) were binocular and equally driven by both eyes. Only one neuron was found to be binocular with a preference for the ipsilateral eye. # = 42
DU
|
^m
7V=118
1 2 3 4 5 Contralateral «- equal -+ ipsilateral
(a) (b) Figure 3. Directional responses and ocular dominance in superior colliculus neurons from darkreared kittens, (a) Distribution of 42 neurons from the 'dark-reared group', (b) Redrawn from the results of Hoffmann and Sherman (1975). Distribution of 118 neurons recorded in ten-monthold kittens lid-sutured at birth. Legend as in figure 1.
518
J-M Flandrin, M Jeannerod
Optokinetic nystagmus. As already mentioned by Vital-Durand et al (1974) OKN is present in visually naive ten-fifteen-week-old kittens, when they are first put in the testing situation. This result, confirmed by the present study, is controversial to that of Van Hof-Van Duin (1976). This author, using a visual inspection technique to assess eye movements, found that five to nine days spent in a normally lighted environment were necessary for OKN to appear in four-month-old kittens dark-reared from birth. In our dark-reared animals, however, OKN was far from normal. Saccades were large, and somewhat erratically distributed in time. The main difference with respect to normal controls was that beat frequency was not positively correlated with stimulus velocity. Instead, the frequency of beats steeply decreased when velocity was increased (figure 5). In two kittens from the 'recovery group', OKN was studied up to the third week after the end of the dark period. No improvement was observed.
W=101
#=101
l
2
1 2 3 4 5 Contralateral *- equal -> ipsilateral
3
(a) (b) Figure 4. Directional responses and ocular dominance in superior colliculus neurons after five week 'recovery' period from dark rearing. Same legend as in figure 1.
40
a*
2 30
* -\V
20
Velocity (° s"1)
50
Figure 5. Optokinetic responses in two dark-reared kittens, recorded in the binocular condition. Same legend as in figure 2.
3.3 Influence of rearing in stroboscopic light Superior colliculus (110 neurons). In the 110 neurons studied from the 'strobe group', the spontaneous activity was low. Eighty-nine neurons (81%) had no response to flashed diffuse light; the remaining 21 (19%) had a weak response. The results of the directionality tests are shown in figure 6a. Only two direction-selective neurons
Motion detection mechanisms in the kitten
519
were found; 11 (10%) had a direction preference, and the majority (97, or 88%) were nondirectional. The results of the ocular dominance test, illustrated in figure 6b, show that 68 neurons (62%) were monocular and driven only by the contralateral eye; 14 (13%) were binocular, with a greater contribution from the contralateral eye; and 27 (24%) were binocular and equally driven by both eyes. Only one neuron was found to be binocular, with a greater tendency to be driven by the ipsilateral eye. These results show that the collicular neurons of strobe-reared cats, i.e. with no experience of visual motion, have directionality and binocularity characteristics similar to those of dark-reared and binocularly sutured cats. Optokinetic nystagmus. OKN was even weaker in the 'strobe group' than in the 'dark-reared group', in the sense that it was less readily obtained when the animals were first placed in the testing situation. OKN displayed the same characteristic as in the 'dark-reared group', i.e., the frequency of beats decreased when the velocity of the stimulus increased. 7V=110
JV=110
l
2
3
1 2 3 4 5 Contralateral
