THE EFFECTS OF ADDITION AND SUBTRACTION OF EYES ON LEARNING IN SALAMANDER LARVAE (Amblystoma punctatum)

Department of Psychology, Indiana University of Pennsylvania, Indiana, Pennsylvania 15701, USA and School of Optometry, Indiana University, Bloomington, Indiana 47405, USA

INTRODUCTION

MATERIALS, METHODS AND EXPERIMENTS

Experiments fell into one of the following major groups:

1. Normals, animals with both natural eyes intact (14 cases);
2. Minus, animals with the left natural eye removed (8 cases);
3. Triclops, larvae with both natural eyes intact plus a supernumerary eye (12 cases):
   

4. Cyclops, with a transplanted eye in the same location as the Triclops but lacking in both natural eyes (7 cases):
   

5. Eyeless (8 cases).

Extirpations of eyes and the transplantations were performed under narcosis during the same interval. The anesthetic was MS 222 (Sandoz) (tricane methanesulfonate) (1: 8000) in spring water. The prospective host region for the eye transplants in Cyclops and Triclops was the epiphyseal portion of the diencephalon exposed by cutting a triangular window in the overlying membranous investing neurocranium. Immediately after producing the window the left eye of a donor was excised and floated into position where it was secured gently but snugly by a simple graft cover. Animals were kept immobile for approximately 4 h. Twenty-one of 24 transplanted eyes survived and became an integral part of the heterotopic site.

The behavioral part of the experiment was initiated 31 days after removal or transplantation of eyes. The training apparatus consisted of a 4 cm wide alley equipped with platinum electrodes. A mild shock (10 V, DC, 10 c/sec) was delivered using pulse and wave form generators (Tektronix types 161 and 162, respectively). The alley was filled with enough spring water to permit an animal to swim freely forward or backwards. Photic stimulation was provided by an American Optical Company Spencer spot lamp set at 5 V, DC. Time intervals were measured with a Kodak timer. Procedures were carried out in a dark room with an ambient temperature maintained at 15 deg. C

Animals were trained in an avoidance paradigm with light as the conditioned stimulus (CS) and shock as the unconditioned stimulus (US). The CS-US interval was 10 sec and the intertrial interval was variable, 10-25 sec. Prior to training, each animal's operant level was established by twenty trials with light alone. Training consisted of fifty trials a day for 4 days (acquisition). On the 5th day, animals were given fifty trials with light alone (extinction). Immediately after acquisition, Normals and Triclops animals were subdivided equally into A and B subgroups. Those designated A were left intact during extinction. The left eye of each Normal-B was extirpated. Both natural eyes of the Triclops-B animals were removed. These procedures were initiated to find out if the reduction in eyes after acquisition would influence extinction.

The functionality of heterotopic eyes was studied following extinction. After extinction both natural eyes were removed from the Triclops-A. These animals, the Triclops-B and the Cyclops were given 50 avoidance trials with careful attention given the quality and patterns of movement in response to light. There are certain characteristics to the normal response to light: immobilization at the immediate onset of stimulus, alerting in response to light, backward movement out of the light. Eyeless animals do not exhibit these responses and when shocked may even swim around the bowl and position themselves in the light. Animals that can see do not do this. All of the Cyclops animals exhibited the normal signs of response to light. Of 14 Triclops originally in the study 12 exhibited typical responses to light after removal of natural eyes; only these 12 are considered in this report. Following the latter observations the eyes of the 12 Triclops and of 4 Normals were removed; this obliterated avoidance. Cyclops were saved for morphological studies to be conducted at a later date.

RESULTS

Fig. 4

Table III contains statistical data performed on the last block of 25 trials. Table IVsummarizes extinction and contains the means and standard deviations for the last 25 acquisition trials. Table V is a summary of calculations to determine the characteristics of acquisition; these results as well as those for rates and acceleration were based upon plots generated between blocks of 50; i.e., session to session. The latter curves do not contain the apparent jump discontinuities that arise when the function is generated between blocks of 25.

Table IV: EXTINCTION: means and standard deviations (+/-)

The left natural eye was removed from animals in Normal-B prior to extinction. Normal-A animals were left intact. Both natural eyes were removed from Triclops-B subjects prior to extinction trials. Animals in Triclops-A group were intact during extinction.

Experiment	Avoidance*			Decrement (Rate)
	I	II	III
Minus	13.1 +/- 2.2	7.1 +/- 3.1	3.0 +/- 2.7	0.210
Normal-A	15.5 +/- 2.9	8.0 +/- 2.8	3.2 +/- 2.1	0.246
Normal-B	15.5 +/- 2.9	9.4 +/- 2.8	5.4 +/- 2.3	0.210
Triclops-A	17.6 +/- 2.0	12.0 +/- 3.4	6.0 +/- 4.6	0.232
Triclops-B	17.6 +/- 2.0	9.0 +/- 3.0	4.8 +/- 5.3	0.256
Cyclops	21.0 +/- 3.2	14.7 +/- 3.6	9.0 +/- 2.9	0.240

*I is aquisition in the last block of 25 trials; II and III refer to blocks of 25 extinction trials each.

Eyeless animals could not be trained (Fig. 3).

The Minus group by the last block of 25 trials were avoiding at the 50% level. Their rate of acquisition was 0.085 with an acceleration of approximately 0.2 (see Table I and Table II).

Normals acquired at a rate of 0.112 with an acceleration of 0.304 and by the final block of 25 trials were performing successfully some 60 % of the time.

Triclops acquired the task at a rate of 0.126 with 0.35 acceleration and a 70% avoidance by the last block of 25 trials.

In the case of Cyclops with an acquisition rate of 0.188 and 0.5228 acceleration, the mean avoidance level was approximately 85 % by the final block of 25 trials.

Acquisition data (excluding Eyeless) were analyzed for the last 25 trials by the Kruskal-Wallis (8) analysis of variance; an H value of 93 (P 0.001) was obtained. Group comparisons were made for the last 25 trials using the one-tailed Mann-Whitney U-test (see Table II) [8]. Significant differences prevailed in each comparison except for Normal versus Triclops.

Extinctions are shown on a percentile basis in Figs. 3 and 4 and are tabulated in two blocks of 25 trials each in Table IV. The decrement for each group was linear and of about the same instantaneous rate for all groups including Normals-B and Triclops-B (see Table IV). The absolute level to which performance fell in extinction was roughly proportional to the heights achieved in acquisition. It seems worthy of special note that the subtraction of one eye from the Normals or two eyes from the Triclops had no significant effect on extinction, in rate or absolute value.

Inspection of the values between blocks of 50 trials suggested that acquisition was logarithmic. If this were indeed true it was reasoned that the absolute value produced in logarithmic integration should, if divided by the absolute value produced in linear integration, yield a quotient of 2.7813 . . . (e). Furthermore, the difference between predicted (e) and observed quotients would yield information about: (a) the extent to which the several curves taken as a family varied from true logarithmic; (b) the deviation of individual curves from the true logarithmic; and (c) the degeneracy of a given curve from the family of curves. These values are tabulated (Table V) as family and functional errors.

Minus, Normal and Triclops curves belong to the same family (see Table V) and the family as a whole is very nearly perfectly logarithmic. The functional characteristics for the Cyclops group varied from the predicted function by 1.6%. This is not wholly insignificant despite its superficially small magnitude. To illustrate, if we were to round off all values to two places the F(A)/F(B) functions in Table V for Cyclops would be 2.8 while those of other groups and the predicted function would be 2.7. This means that in an infinite expansion by

(1 + 1/n)ⁿ

we begin to acquire a perturbation of pure continuity with n somewhere between 2 and 3. Because the functional error remains small (2.8--2.7 = 0.1/2.7 = 3.7 %) after rounding off it seems improbable that the function, per se, is generating the error. Rather, it would appear that at the levels of performance reached by the Cyclops we begin to detect hidden factors that would relate to acquisition in the form of constants or (more likely) partial derivatives of trials. This seems especially worthy of note, for it is quite possible that the Cyclops curve is just beginning to signal the establishment of the task on a long term basis.

DISCUSSION

However, the Cyclops permits us to stipulate the ideal weight of a supernumerary eye. These animals learned the task at an amazing rate, far better than any other group. Clearly the heterotopic eye did not have the same numerical relationship to acquisition as a single natural eye. With Cyclops the acceleration was some 0.5 and it is this value that we must employ to predict the performance of Triclops. The Triclops did not even remotely approach the predicted level of 0.85. Constraints must have been imposed centrally in the Triclops, either on an anatomical or physiological basis or both; i.e., it is possible that ingrowing supernumerary optic nerve fibers in the Triclops were limited in the number of functionally meaningful connections they could make, being pre-empted, perhaps, by terminations of the already present optic nerves of the two intact natural eyes. On the other hand, the primary connections made in the Triclops may have equaled those in the Cyclops, but the additional information carried centrally may have been dampened down and brought algebraically into the net result by inhibition generated in the natural two-eyed state. In either case the central nervous system appears to have played the major role in setting the level of performance. The remarkable thing about this is that it did so with almost integrative precision. In terms of net effect on behavior, the supernumerary eye of the Triclops was made to behave as a normal eye.

There is an intriguing aspect of the data from the Cyclops that further amplifies the constraining capacity of the central nervous system. Cyclops values are close to four times the difference between Normal and Minus; i.e., the undampened supernumerary eye had the effect of four additional natural eyes. In terms of the chosen learning paradigm an eye positioned in the vicinity of the epiphysis is transcendentally more efficient than one bearing an immediate topographic relationship to the lateral diencephalon. We are not suggesting that total 'intelligence' of the animal was enhanced by heterotopic innervation. Indeed, teleologically, such a relationship to optic stimulation might reduce the capacity to learn other things and, perhaps, subtract from the chances of survival. It is intriguing to us that some function of the status quo in optic innervation was responsible for a net--but orderly--decrease in total utilization of the added sensory carrying capacity.

SUMMARY

REFERENCES

1. ALLPORT, F. H., Theories of Perception and the Concept of Structure, Wiley, New York, 1955.

2. GROSSER, O-J., AND CREUTZFELDT, O., Neuronal physiology in the visual system. In M.A.B. BRAZIER (Ed.), Brain and Behavior, American Institute of Biological Sciences, Washington, D.C., 1961, pp. 308-357.

3. LINDSLEY, D. B., Electrophysiology of the visual system and its relation to perceptual phenomena. In M. A. B. BRAZIER (Ed.), Brain and Behavior, American Institute of Biological Sciences, Washington, D.C., 1961, pp. 359-392.

4. ROSNER, B. S., Neural factors limiting cutaneous spatiotemporal discriminations. In W. A. ROSENBLITH (Ed.), Sensory Communications M.l.T. Press, Cambridge, Mass., 1961, pp. 725-736.

5. SCHWARTZ, A. S., AND LINDSLEY, D. B., Critical flicker frequency and photic following in the cat, Bol. Inst. Estud. med. biol. (Mex.), 22 (1961) 249 262.

6. WALKER, A. E., WOOLF, J. 1., HOLSTEAD, W. C., AND CASE, T. J., Mechanisms of temporal fusion. Effect of photic stimulation on electrical activity of visual structures, J. Neurophysiol., 6 (1943) 213 219.

7. SPERRY, R. W., Mechanisms of neural maturation. In S. S. STEVENS (Ed.), Handbook of Experimental Psychology, New York, 1951, pp. 236-280.

8. SIEGEL, S., Nonparametric Statistics for the Behavioral Sciences, McGraw-Hill, New York, 1956, pp. 116127, 184-193.

return to Eyes and Eyes menu

return to Shufflebrain menu

For a non-technical account of this subject go here

Comments or inquiries: pietsch@indiana.edu