Neurotoxicologyand Teratology,Vol. 14, pp. 235-245, 1992

0892-0362/92$5.00 + .00 Copyright©1992PergamonPressLtd.

Printedin the U.S.A. All rightsreserved.

Behavioral Effects of Lead in Monkeys Tested During Infancy and Adulthood D E B O R A H C. R I C E

Toxicology Research Division, Bureau o f Chemical Safety, Food Directorate, Health Protection Branch, Health and Welfare Canada, Banting Building, Tunney's Pasture, Ottawa, Ontario, Canada, K1A 01,2 Received 5 July 1990; Accepted M a r c h 1992 RICE, D. C. Behavioral effects of lead in monkeys tested during infancy and adulthood. NEUROTOXICOL. TERATOL. 14(4) 235-245, 1992.-A total of 12 monkeys (Macacafascicularis) were dosed orally from birth with 0 or 2000 #g/ kg/day of lead as lead acetate. Blood lead concentrations of treated monkeys peaked at an average of 115 t~g/dl by 100 days of age and decreased to a steady state level of 33 ~,g/dl after withdrawal of infant formula at 270 days of age. At 5-6 months of age, they were tested on a nonspatial discrimination reversal paradigm. At 2.5-3.0 years of age, they were tested on a series of nonspatial discrimination reversal problems, including irrelevant cues. As adults, performance was assessed on a differential reinforcement of low rate (DRL) schedule of reinforcement, a spatial delayed alternation task, and during training on a visual discrimination task for a visual psychophysicsexperiment. There were no or marginal deficits on the discrimination reversal task during infancy. Although lead-treated monkeys were impaired on this task as juveniles, they were less impaired than would have been predicted based on their history of blood lead concentrations. Treated monkeys exhibited decreased interresponse times and a greater ratio of responses per reinforcement on the DRL schedule compared to controls. Four of five treated monkeys were unable to learn the visual discrimination task without a remedial training procedure in which the relevant visual stimuli were arranged to appear as if they were on the response buttons. Treated monkeys were unimpaired on the delayed spatial alternation task. The results are interpreted as suggestiveof an interaction between the behavioral history of the monkeys as infants with the results of later behavioral testing. Lead and behavior Behavioreffects of lead in the monkey DRL Delayedalternation Discrimination reversal

RESEARCH over the last decade has clearly established that lead exposure during development in the monkey produces consistent behavioral impairment. Tasks found to be sensitive to lead-induced impairment include nonspatial discrimination reversal (1,26,32,38), spatial discrimination reversal (1,2, 8,28), spatial and nonspatial matching to sample (25), spatial delayed alternation (14,15,33,37), learning set (16), Hamilton search task (13) and fixed interval and fixed ratio (19,27,36), and DRL (31) schedules of reinforcement. In previous experiments in our laboratory, monkeys were dosed with lead continuously from birth, with behavioral testing beginning at about 3 years of age. In the present study, behavioral testing began at 60 days of age in order to determine whether behavioral effects were already present after only 60 days of dosing. The tests chosen had been determined in this laboratory to be sensitive to lead-induced impairment in juvenile or adult monkeys: namely, fixed interval and fixed ratio schedules of reinforcement (36) and nonspatial discrimination reversal (26,32,38). Performance on these tasks was reassessed when these monkeys were juveniles, in order to determine the effects of a relatively high dose of lead on the same tasks and at the same age as in previous groups of monkeys receiving lower doses of lead. The results of performance on a multiple fixed interval-fixed ratio schedule indicated a greater lead-induced behavioral effect in this group of mon-

Attenuation of behavioral effects by early experience

keys compared to groups exposed to lower doses of lead (27). When these monkeys were adults, they were tested on additional tasks which had proven sensitive to lead-induced impairment in our laboratory: DRL (31) and spatial delayed alternation (33,37). Performance was also assessed on a visual discrimination problem during training for a visual psychophysics experiment, in which impairment had not been observed in monkeys exposed to a lower dose of lead (unpublished). Assessment of performance on these tasks allowed further comparison of the performance of this group of monkeys with other groups not tested during infancy on additional learning tasks and an intermittent schedule of reinforcement. The results of these experiments are presented here. METHOD

Subjects and Dosing The dosing regimens and blood lead histories of this group of monkeys has been described previously (27). Monkeys were born of undosed females (Macaca fascicularis) bred in the neurotoxicology primate colony, separated from their mothers at birth, and reared according to standard nursery procedure. They were dosed from birth with 0 (4 males; 2 females) or 2000 (3 of each sex) #g/kg/day of lead as lead acetate. Blood lead concentrations of treated monkeys peaked at 115 ttg/dL 235

236 by 100 days of age, and decreased to steady state levels of 33 # g / d L following withdrawal from infant formula at 270 days of age. Blood lead values of controls were 3-4 #g/dL.

Behavioral Methods Apparatus and behavioral history. When tested as infants, monkeys were housed individually in stainless steel cages, with an open front that allowed fronts of different types to be attached (24). The behavioral equipment was mounted on a frame outside the cage. The feeder was a solenoid valve attached to a stainless steel tube through which infant formula (50o70 SMA, Wyeth Ltd. North York, Ontario) was delivered as a reinforcer. Pushbuttons on either side of the feeder were clear plastic pigeon buttons, requiring less than 15 gm to close. These could be backlit with colors a n d / o r two dimensional white forms. Beginning at 60 days of age, monkeys were shaped to press the buttons, and were then tested on a fixed ratio (FR) schedule of reinforcement (27). Following testing on the FR, they were tested on a nonspatial form discrimination and series of reversals, described in this article, beginning at 5-6 months of age. Infants were then tested on a chain F R fLxed interval (FI) schedule (27), which was the final schedule tested during infancy. Monkeys were tested 7 days per week, 16 h per day, beginning at 3:00 p.m. This regimen results in no caloric deprivation, and allows infants to grow at a normal rate (24,27). When the monkeys reached 2.5-3.0 years of age, behavioral testing was recommenced. Performance was first examined on a series of discrimination reversal tasks, described in this article. This was followed at 3.0-3.5 years of age by assessment of performance on a multiple F I - F R schedule, described previously (27). They were then not tested until 7.07.5 years of age, at which time they performed on a differential reinforcement of low rate (DRL) schedule, followed immediately by a spatial delayed alternation task. For all tasks, monkeys sat in a primate chair, restrained only at the neck. They faced a clear Plexiglas panel with a stainless steel tube for delivery of the apple-juice reinforcer. The response manipulanda were two clear plastic pushbuttons that could be backlit with colors or forms centered on either side of the feeder tube. The behavioral equipment was housed in a soundattenuating chamber. House light and white noise were on for all sessions. Monkeys were tested 5 days per week. When these monkeys were 9.5-10.0 years old, they began training on a visual discrimination task in preparation for a visual psychophysics experiment. The apparatus is described under specific methods. When these monkeys were tested as infants, one treated monkey was eliminated from testing because he failed to progress through the training procedure to press the button as described previously. When these monkeys were tested as juveniles, this individual and a behaviorally naive control monkey were added to the experiment. Schedule control and data acquisition in all experiments were by means of a Nova minicomputer (Data General Corp., Southboro, MA) utilizing a behavioral notation system developed in the laboratory (7). All data were stored as uniquelycoded interevent times so that the session could be completely reconstructed from the raw data. Discrimination reversal-infants. For the initial discrimination training, red was the positive stimulus and a dark button the negative (training step 1). The position (left or right) of the positive stimulus was chosen randomly from trial to trial. A correction procedure was used; an incorrect response

RICE was followed by a repeat of the same trial. Incorrect responses resulted in a 10 s time-out period, during which the buttons were dark. A session consisted of one dally 16 h test period. When the monkey responded with an accuracy of 90070 or better for 1 session, a cross was substituted for red as the positive stimulus (training step 2). After 2 sessions, a triangle was introduced on the other button as the negative stimulus (initial acquisition). A series of 9 discrimination reversals was then instituted, in which the positive stimulus became the negative stimulus, and vice versa. For each reversal, criterion for advancement was 87°7o correct or better for 5 (not necessarily consecutive) sessions. Data analysis was performed on blocks of trials after a specific change in schedule (i.e., change in stimuli during training or institution of a reversal). This strategy was adopted for a couple of reasons: (A) Because session length was defined by time rather than by number of trials, analysis of data by session would include different numbers of trials within a session, and session 2 of a reversal for one monkey might be several hundred more trials into the reversal than for another monkey, for example. Analysis by trials eliminated that problem; (B) It became clear as the experiment progressed that the criteria for change from one training step to another or for a reversal were extremely lenient. Monkeys generally performed above 80070 correct by 100-200 trials. It is generally the case that lead-induced deficits are manifested most clearly immediately following a change in schedule contingencies (8,26,32). Therefore, analysis focused on early trials after a schedule change (trials 1-10, 1-50, 51-100, and 101-200). In addition, asymptotic performance was assessed by examining per cent correct responses for trials 501-1000, to determine whether this differed between treated and control monkeys. The following parameters were examined for each training step, the initial acquisition, and each reversal: 1. The per cent correct, was calculated for trials 1-I0, 1-50, 51-100, 101-200, and 501-1000, where a trial included all correct and incorrect responses between reinforcements. 2. For trials 1-50 and 51-100, per cent correct on switch and nonswitch trials was determined. A switch trial was one in which the positive stimulus was on the button opposite from that o f the previous trial, and a nonswitch trial was one in which the positive was on the same button as in the previous trial. 3. For trials 1-50 and 51-100, the perseveration ratio was calculated, defined as the number of perseverative errors divided by the total number of errors. A perseverative error was defined as any error(s) following the first incorrect response of any trial. 4. The number o f responses required to reach a criterion o f 8707o correct was calculated.

Discrimination reversal-juveniles. The discrimination reversal task was identical to that used previously (26,32). Although the exact task differed from that used during infancy, it represents an age-appropriate adjustment in task difficulty. The experiment consisted of three tasks: Task 1, a form discrimination (white square vs. white triangle, both with a red surround); Task 2, a color (red vs. green) discrimination with irrelevant form (square vs. triangle) cues; Task 3, irrelevant form discrimination with irrelevant color cues, using the same stimuli as in Task 2. For tasks 2 and 3, all 4 of the form-color stimulus combinations were presented. Each session was considered to be divided into ten 10-trial blocks. For each task, when the monkey made 0 or 1 error in a block, a reversal was instituted (i.e., the positive stimulus became the negative and

LEAD-INDUCED BEHAVIORAL EFFECTS vice versa). A total of 15 reversals plus the initial acquisition were run for each task. There was no correction procedure for errors. A correct response resulted in apple-juice reinforcement; an incorrect response resulted in a seven second time out period. Trials were separated by a 3-s intertrial interval. The training procedure consisted of one session of continuous reinforcement on each button lit with blue; criterion for advancement to the next stage was 100 correct responses in 0.5 h. This was followed by a blue (positive)-yellow (negative) discrimination training procedure; criterion for advancement to Task 1 was a total of 3 sessions with fewer than 5 errors over the 100 trials per session. The number of errors was totalled for the initial acquisition and each reversal for each of the 3 tasks. Attention to irrelevant cues was examined as described previously (26,32). The ratio of incorrect responses on each button, and for Tasks 2 and 3, the ratio of incorrect responses on irrelevant stimuli, were calculated by number incorrect on button (or stimulus) number incorrect on button (or stimulus) + number correct on button (or stimulus) The ratio of trials on which a monkey avoided a particular button or irrelevant stimulus was calculated by number incorrect on other button (or stimulus) number correct on other button (or stimulus) + number correct on button (or stimulus) The totals of the larger ratio for each pair (i.e., right incorrect and left incorrect; triangle avoid and square avoid, etc.) over the course of each task were calculated to circumvent the fact that different individuals may show preference for different irrelevant cues. A total measure of attention to irrelevant cues was calculated by adding these latter values. These analyses were performed for each task by totalling the values for the acquisition plus all reversals. Differential reinforcement of low rate (DRL). The DRL schedule was identical to that described previously (31). The terminal schedule was a DRL 30 s, preceded by two sessions of DRL 5 s, and 1 session each of DRL 10 and 20 s. The button on the right side of the feeder was lit with blue to indicate the schedule was in effect. The monkey was required to withhold responding for at least the length of the DRL value to be reinforced. A response longer than this interval after either the preceding (nonreinforced) response or reinforcement was reinforced. Responding before the DRL value elapsed reset the contingency and the monkey had to wait a further length of the DRL value for an opportunity for reinforcement. Sessions were 45 min long. A total of 60 DRL 30 s sessions were run. Data were analyzed as follows: 1. Mean and median interresponse time (IRT) (the time between successive responses or between a reinforcement and a response). 2. Number of reinforced responses. 3. Number of nonreinforced responses. 4. The ratio of nonreinforced to reinforced responses (responses/reinforcement).

Delayed alternation. The behavioral methodology was identical to that described previously (33,37). The training schedule consisted of a tracking procedure. For the first trial of a session, both buttons were lit yellow. Response on either

237 button resulted in reinforcement. For subsequent trials, the light appeared on the button opposite the one on which a response was last reinforced; responding on the lit button resulted in reinforcement. There was a 100 msec delay between trials, during which both buttons were unlit. The schedule contained a correction procedure; if the monkey responded on the wrong button, the trial was repeated and the total number of trials in the session was increased by one. Responding during the delay reset the delay to its initial value. Sessions were 100 correct trials. Criterion for advancement to the next schedule was 5 sessions with 5 or fewer errors, or a maximum of 20 sessions. For the alternation schedules both buttons were lit. On the first trial of the session, the monkey could respond on either button to be reinforced. On subsequent trials, the monkey was required to alternate responses between buttons. The first alternation schedule contained a 100 msec delay between trials. Criterion for advancement was 5 or fewer incorrect responses for 5 sessions, or a maximum of 20 sessions. A series of sessions with increasing delay values was then instituted: 0.5 s (10 or fewer errors for 5 sessions or a maximum of 20 sessions), 1.0 s (10 sessions total), 3.0 s (10 sessions total), 5.0 s (20 sessions total), and 15.0 s (20 sessions total). Data analysis was performed for the group on the least number of sessions completed by any monkey for each delay value. The total number of incorrect responses and the number of responses during the delay period were calculated. Training for a visual psychophysical experiment. Five of the 6 lead-treated and 4 of the control monkeys from the same group were trained for assessment of visual function. The sixth treated monkey was not tested due to refractive errors in both eyes, and only 4 control monkeys were required to provide normative data. The monkey sat in a sound-attenuated room painted flat black facing two oscilloscopes. At about the level of the monkey's waist were two pushbuttons, one on either side of a feeder tube mounted at an angle so that it did not obstruct the monkey's view of the oscilloscopes. During training, one oscilloscope displayed a blinking vertical square wave grating (i.e., bars) while the other contained a blank field of equal average luminance. The scope displaying the grating varied randomly between trials. The monkey's task was to press the button corresponding to the scope displaying the grating in order to be reinforced. A correction procedure was used; an incorrect response resulted in the positive stimulus appearing on the same scope for the next trial. Trials were separated by a 3-s intertrial interval; an incorrect response resulted in a 7-s time-out period. After the monkey learned the task, the schedule requirement was changed so that two responses on the same button were required to complete a trial (FR2), followed by the terminal schedule of three required responses (FR3). Initially, the scopes were about 2 feet from the monkey's eyes. They were moved back gradually over successive sessions, depending on the performance of the monkey, to the 57 inches required for testing. The blinking of the scope containing the grating was then eliminated, the square wave was changed to sine wave, and psychophysical testing was begun. After 5 weeks of training, some treated monkeys were still responding randomly with the scopes 2 feet away, with no indication of learning the task. A remedial training procedure was therefore implemented. This consisted of introducing a Plexiglas sheet on which were mounted pushbuttons at the level of the monkey's eyes, a conformation with which they were familiar. The scopes were placed directly behind these buttons. Initially, the Plexiglas was blocked with cardboard

238

RICE

so that the scopes were visible only through the buttons. When the monkey was responding with few errors, the cardboard was removed and the scopes were moved back over successive session to 24 inches. The scopes were then moved close again; the special apparatus removed, and the regular buttons became the response manipulanda. The scopes were then moved back over successive sessions to the required 57 inches. One treated monkey failed to make the transition from the remedial eye-height buttons to the regular waist-height buttons. For this individual a Plexiglas front with slots allowing the buttons to be moved down gradually to the final required position was introduced. The buttons were gradually moved down over successive sessions and then the special apparatus removed when these buttons were directly aligned with the buttons at the required position. Statistical analysis. For statistical analysis of the infant discrimination reversal data, data were separated into 2 parts: (A) the training steps and the initial acquisition and (B) reversals 1-9. The data for training steps and initial acquisition were not characterized by any particular pattern, whereas the data for the reversals represented an orderly learning curve. As well, the reversal data represent repeated measures of the same experimental procedures, whereas the training steps are not logically connected in the same way. Therefore, data were analyzed separately for each of the two training steps and the initial acquisition. For the reversals, orthogonal polynomials were fitted to most measures, with reversal treated as a time variable (20). For the number of trials required to reach 87°7o, a negative exponential model, y = t~e-2~ was fitted. This analysis provided a better fit for these data than the polynomial analysis. The resulting coefficients were analyzed by t statistic using a randomization test (6). The randomization test calculated the p value by repeatedly randomizing the values of the treated and control group into two groups of equal size and determining how unusual the observed difference was. In these analyses, all possible rearrangements of the data were generated. The randomization procedure may be used to generate a p value for any choice of a test statistic, without regard to the usual assumptions for

parametric statistical analyses. This statistic was chosen because past experience indicated that group variance is often greater in lead-treated groups than control, with extreme outliers not unusual. For the analysis of the discrimination reversal as juveniles, DRL, and delayed alternation experiments, randomization tests were used. Because these behavioral tasks had been assessed several times in lead-exposed monkeys in our laboratory, specific hypotheses about direction of effect could be postulated. Therefore one-tailed tests were used, assuming treated monkeys would perform more poorly than controls, exhibit more attention to irrelevant cues, and perform the DRL with lower IRTs and therefore more nonreinforced and fewer reinforced responses. One-tailed analyses were used in previous assessments of lead-treated monkeys on these tasks (32,33). For each of the discrimination reversal tasks, reversals 1-15 were compared as a block, as were reversals 1-5, 6-10, and 11-15, as has been done previously (26, 32). Acquisition was also analysed for each task. Statistical analysis was performed for total attention to irrelevant cues for each task; if that analysis was significant (p < .10) each stimulus was analyzed separately. For the delayed alternation task, data were analyzed by delay value. For the DRL, the first session (DRL5) was analyzed separately, because these data comprise the initial response to a DRL schedule. All 60 DRL 30 sessions were analyzed together. For the attention to irrelevant cues, p < 0.10 was used as a trigger for analyzing individual stimuli because results for the total measure could obscure significant differences on individual stimuli. For all other comparisons, p < 0.05 was considered significant. RESULTS Discrimination R e v e r s a l - I n f a n t s

Monkeys completed between 300 and 900 trials per session on average. The number of triais required to reach the 870/0 acquisition criterion was about 200 for both groups (Table 1) and monkeys were responding at 80°/o correct by trials 100-

TABLE 1 PERFORMANCE DURING INFANCYON REVERSALS I-9 OF A NONSPATIALREVERSAL DISCRIMINATIONTASK Meanresponse Percentcorrect for Trials 1-10 Trials 1-50

Trials 51-100

Trials 101-200 Trials 501-1000

Overall Overall Switch Non switch Overall Switch Non switch Overall Overall

Control

Treated

42.1 58.7 51.1 83.5 76.3 65.0 91.7 81.7 88.7

p value" Intercept

Linear

38.3 53.8 47.7 79.2 74.6 65.8 90.7 80.4 89.1

.56 .44 .72 .15 .75 .94 .61 .88 .89

.61 .81 .78 .63 .52 .38 .94 .86 .77

.62 .36 214

.05 .21 .64

.33 .15 .65

Perseveration ratio for:

Trials 1-50 Trials 51-1(30 Number o f trials to reach 87% correctb

.53 .27 211

hwo-sided p values from randomization test on model parameter estimates, bIntercept and linear refer to parameter estimates for t~ and fl, respectively from the negative exponential model.

LEAD-INDUCED BEHAVIORAL EFFECTS

239 button. This is undoubtedly the result of the fact that most of these monkeys are right-handed for button pushing; 1 control and 1 treated monkey are left-handed. Therefore position bias would more likely be on the right. Treated monkeys tended to attend to irrelevant form stimuli more than controls for all four measures, which was significant at t h e p = 0.05 level for incorrect responses on the triangle.

200 following a reversal. Monkeys therefore reached the reversal criterion within the first session, and the following 4 sessions (days of testing) represented hundreds or thousands of trials of overtraining. For most measures there was no difference in performance between treated and control monkeys. There were no differences for any of the measures for the training steps or the acquisition of the task. For the analysis of the reversal data, only the intercept of the perseveration ratio for trials 1-50 was different between groups, being higher for the treated than the control monkeys. Thus, even this relatively exhaustive analysis failed to reveal strong evidence of impairment of nonspatial discrimination reversal performance in these monkeys as infants.

DRL Treated monkeys performed differently from controls on the DRL schedule (Table 4). Over all 60 DRL30 sessions, treated monkeys had lower mean and median IRTs (Fig. 3), more nonreinforced responses, and a higher ratio of responses per reinforcement (Fig. 4). The mean IRT for control monkeys was 32 s, longer than the DRL value, whereas for the treated monkeys it was 25 s, below the IRT required for reinforcement. There were no differences between the groups for the first 5 DRL sessions, indicating that the differences observed in the later sessions were a result of the schedule contingencies imposed by the DRL schedule.

Discrimination Reversal-- Juveniles Treated monkeys were not impaired on the acquisition of any of the three tasks (Table 2). Lead-treated monkeys were not impaired on Task 1, the form discrimination with no irrelevant cues. On Task 2, the color discrimination with irrelevant form cues, treated monkeys were impaired over all reversals (Fig. 1). When performance on Task 2 was examined by groups of reversals, treated monkeys were found to be impaired over reversals 1-5. Lead-treated monkeys were not impaired on Task 3, the form discrimination with irrelevant color cues. For the analysis of attention to irrelevant cues, treated monkeys were not different from controls for Task 1 (p = 0.24) or Task 3 (p = 0.13). For Task 2, however, there was an indication that treated monkeys attended to irrelevant cues to a greater degree than controls (/7 = .07) (Table 3). Analysis of individual stimuli revealed that treated monkeys responded on the right button and ignored the left significantly more than control monkeys. There was no such effect for the left

Delayed Alternation There were no differences between groups for the number of sessions required to reach criterion on the tracking procedure, 0.10 or 0.50 s delay schedules. Two treated and 4 control monkeys required the maximum 20 sessions on the 0.10 second delay, although in all cases error rate was less than 10% for at least the last few sessions. One control and 1 treated monkey required all 20 sessions at the 0.50 s delay. Treated monkeys did not perform more poorly than control monkeys as measured by the number of errors across delay values (Table 5, Fig. 5), nor were there differences between groups for the number of responses during the delay.

TABLE 2 NUMBER OF ERRORS ON A SERIES OF DISCRIMINATIONREVERSALTASKS AS JUVENILES Mean(SD) Controls Task 1 Acquisition Reversals 1-15 1-5 6-10 11-15 Task 2 Acquisition Reversals

Task 3 Acquisition Reversals

66.2 34.1 71.7 21.3 9.3

1-15 1-5 6-10 11-15

5.5 6.5 6.5 6.4 6.4

1-15 1-5 6-10 11-15

36.5 18.1 27.3 13.6 13.4

(41.5) (7.8) (17.9) (16.8) (5.0)

Treated

p value'

63.6 35.9 71.7 23.5 12.6

(31.8) (17.1) (32.0) (21.5) (9.3)

0.77 0.20 0.25 0.21 0.1I

(3.2) 5.7 (2.9) 9.5 (3.3) 9.7 (3.2) 8.5 (3.4) 10.2

(2.5) (4.8) (4.1) (5.8) (7.4)

0.25 0.05 0.04 0.12 0.07

(15.1) (5.1) (10.5) (3.9) (7.0)

0.83 0.77 0.25 0.18 0.66

(34.6) (3.4) (5.3) (4.4) (3.1)

aOne-sidedp value from randomization test.

27.7 17.8 27.0 14.5 11.9

240

RICE TABLE 3

80 TASK I

ATTENTION TO IRRELEVANT CUES ON TASK 2 OF A DISCRIMINATION REVERSAL TASK AS JUVENILES

0

60

OA

o&

4C

p value"

20

O'

I

I

I

o

(/) ry 0 ry ry bJ LL 0 rY b.I r~

I

~

20' TASK2

$ 0

O ,t

~

2®,

Z

0

.07 .01 .79 .90 .01 .07 .05 .08 .07

aOne-sided p value from randomization test,

0

10

Total Incorrect on right Incorrect on left Avoid right Avoid left Incorrect on square Incorrect on triangle Avoid square Avoid triangle

A

, 2000

piece of apparatus that allowed the buttons to be slid gradually to the required terminal position, required 19 additional weeks of training. This monkey was also incapable of tolerating the FR3 response requirement, and was tested with only an FRI required for a response choice. Control monkeys and the one treated monkey that did not require remedial training completed the entire training procedure in 20-30 sessions (4-6 weeks). DISCUSSION

40 TASK3

20

#

0

0

2OOO

DOSE (ug/kg/doy) FIG. 1. The number of incorrect responses for monkeys dosed with 0 or 2000 # g / k g / d a y of lead for reversals 1-15 of a series of nonspatial discrimination reversal tasks. Task 1, form discrimination; Task 2, color discrimination with irrelevant form cues; Task 3, form discrimination with irrelevant color cues. Each symbol represents an individual monkey.

Trainingfor Vision Psychophysics Only 1 of the 5 treated monkeys learned the visual discrimination problem without requiring remedial training, whereas all 4 control monkeys learned the task without difficulty. Three of the treated monkeys required about 6-10 weeks to progress through the remedial training (in addition to the initial unsuccessful 5 weeks o f training), to get back to the condition of nonblinking scopes 57 inches from the monkey's eyes (Table 6). The fourth monkey, requiring an additional special

In the present study, monkeys with relatively high blood lead levels were tested during infancy and later in life on various behavioral tasks. Lead-treated monkeys were relatively unimpaired on a simple nonspatial discrimination reversal task as infants, and less impaired on a series of nonspatial discrimination reversal tasks as juveniles than would be expected on the basis of their history of blood lead concentrations. These monkeys were unimpaired on a spatial delayed alternation task as adults. Lead-treated monkeys exhibited differences from controls on a DRL schedule of reinforcement as adults. This group of monkeys was also impaired in their ability to learn a simple visual discrimination task when the stimuli were not directly on the response buttons. The decreased IRT values, increased nonreinforced responses, and increased ratio of responses per reinforcement observed on the DRL schedule is consistent with previous results from these monkeys on intermittent schedules of reinforcement, and extends previous findings on DRL performance in monkeys dosed from birth with lower doses of lead than those in the present study. As infants, lead-treated monkeys in the present study exhibited increased pause times on an FR and decreased pause times on an FI (27). As juveniles, the lead-treated group had increased run rates, as well as increased pause times and index of curvature on an FI as part of a multiple F I - F R schedule. The decreased average IRT in the present study is consistent with the increased rate observed on FI performance. In a previous experiment with monkeys exposed from birth to lower doses of lead and having much lower blood lead levels (31), lead-exposed monkeys were impaired in the acquisition of DRL performance, with no differences in measures of DRL performance after the first few sessions. Thus, the present group of monkeys may be considered to exhibit a greater effect of lead exposure than the groups exposed to lower lead levels. The largely negative results from the discrimination rever-

LEAD-INDUCED BEHAVIORAL EFFECTS

REVERSALS 1-5 20

[]

~o

AO

I

•

I

0

2000

00 FF 30 REVERSALS6-10 O FF nd i,i 20 i, O 0 nd Ld IO m Z

0

0 '

0'

~O0

30 REVERSALS 11-15

20

0

0

0

10 8 ,

DOSE

[3

A

(ug/kg/day)

FIG. 2. The number of incorrect responses for monkeys dosed with 0 or 2000/~g/kg/day of lead for reversals 1-5, 6-10, and 11-15 of Task 2 of a series of nonspatial discrimination reversal tasks. Symbols as in Fig. 1.

sai task during infancy seem surprising, especially in view of the high blood lead levels during testing (115 #g/dL). Moreover, several other groups with much lower lead exposures had exhibited lead-induced impairment on discrimination reversal performance as juveniles or adults (26,32,38). It is unlikely that the duration o f lead exposure had been insufficient to produce behavioral impairment in the lead-exposed monkeys, because they were different from controls on both FR and FI performance during this same period. It is more likely that the task was simply too easy even for infants, particularly with respect to the extremely lenient criteria (several sessions) between reversals. As pointed out in a recent article on testing

241 methods for infant monkeys (34), infant monkeys are capable of easily learning much more difficult tasks and of adapting their behavior successfully to meet much more demanding changes in schedule criteria. It has been demonstrated in this laboratory as well as others (8,26,32,43) that difficult tasks are more sensitive to lead-induced impairment than easy ones. The one parameter on which lead-related impairment was observed, i.e. increased perscverative responding in lead-exposed monkeys, is consistent with other studies from this laboratory (8,25,26,32,33,37). As discussed previously (33), persevcration seems to be a consistent result of lead exposure on a variety of tasks in the monkey. Although lead-treated monkeys were somewhat impaired on the nonspatial reversal task as juveniles, they were less impaired than would have been predicted on the basis of the history of blood lead concentrations. In previous studies in this laboratory, monkeys dosed from birth and having a peak blood lead concentration of 32 p g / d L and a steady state level of 19 p g / d L were impaired over all three tasks (32). Monkeys with peak blood lead levels of 25 /~g/dL were impaired on Tasks 1 and 2, while a group with a peak level of only 15/~g/ dL was impaired only on Task 2 (26). This latter group exhibited the same pattern as the monkeys in the present study: impairment only when irrelevant cues were introduced. Treated monkeys in the present study also attended more to irrelevant cues on Task 2, when irrelevant stimuli were introduced and before they became familiar. This same effect was observed in monkeys as just described with much lower blood lead levels (26), while monkeys with peak blood lead concentrations of 32 p g / d L attended more to irrelevant cues on all three tasks (32). Thus, even this consistently observed effect of lead appears to be attenuated in the present study. In addition, in the present experiment there were not treated individuals with severe impairment as measured by the number of errors, as was the case in previous studies. It is tempting to postulate that exposure to a nonspatiai discrimination reversal task during infancy attenuated any potential lead-induced impairment in these monkeys when they were exposed to a similar series of tasks as juveniles. Treated monkeys were not impaired on the spatial delayed alternation task. This is in contrast to results from monkeys with lower blood lead levels, in which deficits have been consistently observed in this laboratory (33,37). The control group in the present study made more errors than the control groups in previous studies, which was the result of the performance of 2 of the 6 control monkeys. However, the treated monkeys in the present study did not display the marked persevcrative behavior, manifested as repeatedly responding on one button, that was displayed by individuals in previous studits in which the histories of blood lead levels were considerably lower than in the present study (33,37). Lead-treated monkeys in the present study were also not impaired in their ability to learn the task, as had been observed in previous studies in this laboratory. In both previous studies, the behavioral histories of the monkeys were similar to those in the present study beginning from the juvenile period: exposure to intermittent schedules, followed by nonspatial discrimination reversal, followed by the delayed alternation task. However, previous groups were not tested during infancy. Rats exposed to lead beginning post-weaning displayed improved performance on a delayed alternation task (4). However, the training procedure included extensive exposure to a cued alternation task; the authors postulated perseveration of alternating behavior in lead-exposed rats as an explanation for the improved performance. Monkeys at the University of Wisconsin, Madison,

242

RICE TABLE 4 PERFORMANCE ON A DRL SCHEDULE OF REINFORCEMENT DRL30 (60 sessions)

DRL5 (first session) Mean (SD) Controls Nonreinforced responses Reinforced responses Mean IRT Median IRT Responses/reinforcement

395 143 5.2 2.4 3.8

Mean (SD) Treated

(174) (31) (1.9) (0.8) (1.3)

574 134 4.0 2.2 5.2

(361) (25) (1.9) (1.0) (2.6)

p value'

Controls

.08 .15 .07 .IS .07

69 34 32 27 3.9

Treated

(22) (9) (5) (3) (1.7)

96 32 25 23 6.4

p value'

(42) (8) (4) (6) (3.1)

.05 .13 .002 .026 .025

aOne-sided p value based on randomization test.

I000-~CONTROLS

I0001 TREATED

]

v

I

Z

< lO0-

W w (,9

z 0

10-

n

w n/

I-

I

I

I

I

I

I

I

1

10

20

30

40

50

60

lJ I

I

r-

i

I

I

1

1

10

20

30

40

50

60

SESSIONNUMBER

SESSIONNUMBER

FIG. 3. Mean IRT values across the 60 DRL30 second sessions, for control (left) and monkeys exposed to 2000 #g/kg/day of lead. Each symbol represents an individual monkey. Note semi-log scale.

~- I00~ CONTROLS

10o- TREATED

I

Z 10

10-.

\

1

,

I

'

I

I

I

I

1

10

20

30

40

50

60

SESSIONNUMBER

I-

I

I

I

|

I

I

I

1

10

20

30

40

50

60

SESSIONNUMBER

FIG. 4. The number of responses per reinforcement over all 60 DRL30 second sessions for monkeys dosed with 0 or 2000 /~g/kg/day of lead. Symbols and axes as in Fig. 3,

LEAD-INDUCED BEHAVIORAL EFFECTS

243

TABLE 5 PERFORMANCE ON THE DELAYED ALTERNATION TASK Number delay responses

Total number incorrect Mean (SD) Delay value (Scc) Tracking 0.1 0.5 1.0 3.0 5.0 15.0

Mean (SD)

Controls

Treated

p value"

Controls

Treated

9.5 (10.5) 36(25) 31 (30) 22 (15) 37 (33) 35 (35) 54 (40)

14 (15) 22 (11) 14 (15) 15 (ll) 31 (ll) 15 (8) 26 (9)

.15 .93 .93 .88 .80 .96 .96

2.4(4.7) 9.9 (21.1) 31 (39) 10.3 (9.8) 6.8 (9.7) 4.8 (5.3) 10.3 (7.7)

3.6 1.1 9.0 9.1 4.8 3.5 6.8

p value'

(4.7) (1.7) (10.1) (16.2) (8.0) (6.8) (5.1)

.22 .89 .91 .78 .87 .85 .90

aOne-sided p value based on randomization test.

I-

o

150-

W n"

150"

.

IZ

0 0

100"

z 100I

n, W

m

50

50"

Z /

I-

0

t-

O' T

0.10 0.50 1.00 3.00 5.00 15.0

T

0.10

0.50

1.00

3.00

5.00

15.0

DELAYVALUE(sec)

DELAYVALUE(sec)

FIG. 5. The number of errors on a delayed alternation task for control (left) and monkeys dosed with 2000/~g/kg/day of lead. Each symbol represents an individual monkey. T represents the tracking (initial training) task.

TABLE 6 NUMBER OF SESSIONS REQUIRED BY FOUR LEAD-TREATED MONKEYS ON REMEDIAL TRAINING PROCEDURES OF A VISUAL DISCRIMINATION PROBLEM Lower (normal) Buttons, Scopes Blinking 25 25 25 25

Scopes VisibleOnly Through Buttons

Upper Buttons

Lower Buttons, ScopesBlinking

4 2 4 5

14 29 12,4a 40b,23c

14 27 5%16 24

Total Number 57 83 66 117

aMonkey switched back to upper buttons after making many errors on lower ones, then back to lower buttons, bMonkey would not switch to lower buttons, eSpecial apparatus with buttons slid gradually over successive sessions to lower position.

244 exposed to an extensive battery of learning tasks beginning during infancy (see following) were reported to display improved performance on a delayed spatial alternation task (15). However, as discussed in a previous article (33), the main statistical analysis was, in fact, negative, suggesting consistency with the present results. Any sparing of lead-induced behavioral impairment did not generalize to all learning tasks, as exemplified by the difficulty in acquisition of the visual discrimination task when the stimuli were not directly on the response buttons. Of the approximately 40 monkeys trained in the author's laboratory on this task, some of which were behaviorally naive, only monkeys from this treated group exhibited this type of impairment. Other groups included monkeys dosed developmentally with methylmercury (30,35) or a lower dose of lead than in the present study (unpublished). It is well established that naive monkeys learn visual discrimination tasks more slowly if the discriminative stimuli and response manipulanda are separated (see 5,40 for reviews). It may be that the increased task difficulty imposed by the cue-response separation, coupled with the relatively drastic change in a "rule" that was part of the environment of these monkeys from infancy (i.e., that relevant visual stimuli for the task at hand appeared on the response buttons) unmasked lead-induced behavioral impairment in the treated monkeys. A cohort of monkeys at the University of Wisconsin has undergone testing from infancy into adulthood on similar tasks as the monkeys in the present study. Monkeys were dosed with lead from birth to 1 year of age, with peak blood lead levels of 80 or 130 # g / d L during dosing. As infants, they were tested on spatial and nonspatial discrimination reversal problems in a maze, using access to a diaper as reward (1). When tested as 4-year-olds on a spatial discrimination reversal problem in a Wisconsin General Testing Apparatus, treated monkeys were impaired (2). They were also impaired as juveniles on the Hamilton search task, a spatial memory task (13). Treated monkeys were not impaired on a spatial delayed alternation task when tested during adulthood (as just discussed) (15). There are several differences between the Wisconsin study and the present one. The Wisconsin monkeys were tested in a maze as infants but not as juveniles or adults. It may be that the difference in response topography (and/or reinforcer) negated any possible attenuating effect of early experience. This would be consistent with the results from the present study in which treated monkeys were unable to generalize from the condition in which relevant stimuli were displayed on the response buttons to the condition in which they were not. In addition, the infants in the Wisconsin study were not exposed to an extremely lenient reversal criteria as were the monkeys in the present study. This may account for the fact that the Wisconsin monkeys were impaired during infancy. It may also be the case that the monkeys in the Wisconsin study would have exhibited greater impairment as adults had they not been tested during infancy. There are at least 2 potential (not mutually exclusive) explanations for the results obtained in the present experiment. One is that testing during infancy provided a generalized enriched environment sufficient to attenuate the effects of lead; the other is that the behavioral history of the monkeys interacted specifically to affect the outcome of later behavioral testing. It is well established that early environmental enrichment can attenuate a variety of insults (see 9 for review). Early enrichment results in improved performance on a variety of learning tasks in monkeys (11) and rats (41,42). Various forms of en-

RICE richment, including early operant conditioning, also produces changes in such parameters as dendritic branching (10, 12,18,39), which may be adversely affected by lead exposure (17,22,23). In a study in rats (21), an early enriched environment attenuated some behavioral deficits induced by developmental lead exposure. However, as has been hypothesized in the foregoing discussion, it appears that behavioral testing during infancy produced differential effects on subsequent behavioral tests. It seems unlikely, therefore, that early testing produced a global enhancement of performance, but rather that the specific behavioral history is responsible for the pattern of attenuation of impairment observed. The assertion of relative sparing of the present group of monkeys in certain tasks is based on a comparison to results of previous studies, in which monkeys with lower blood lead levels were more impaired than the present group of monkeys. However, although the monkeys in previous studies were not tested during infancy, they were tested on a number of behavioral tasks, the results of which could potentially be influenced by performance on preceding tasks. It appears, then, that confounding behavioral history per se is an insufficient explanation for the results of the present study. The most likely explanation is that these results represent the consequences of the specific behavioral history of these monkeys during a certain period of development. That early behavioral history may have influenced the outcome of later tests raises the issue of the effects of behavioral history in general on lead-induced behavioral impairment. Most of the research in monkeys has involved repeated testing of sets of animals, which could serve to either attenuate or exacerbate lead-induced impairment. Although the data presently available do not allow assessment of this issue, nonetheless, in many cases tasks were performed in a different sequence by different groups, both within and between laborat o r i e s - y e t the effects of lead on similar tasks has been remarkably consistent (29). In addition, the rodent literature is generally in good agreement with the primate literature (3), even though rodents have not generally been tested on series of behavioral tasks. It seems likely, then, that although repeated testing of monkeys has probably added variability to the body of data, it has not severely compromised the interpretation of such studies. It appears from the behavioral tasks examined in these monkeys that performance on intermittent schedules of reinforcement is relatively resistent to modification by early experience. Lead-exposed monkeys performed differently from controls as infants, juveniles, and adults, to a greater degree than previous groups of monkeys exposed to lower doses of lead. Exposure to a reversal task during infancy, on the other hand, seems to have attenuated lead-induced impairment on reversal tasks later in life. This protection did not extend to a different type of learning task, however, where treated monkeys were severely impaired. These data suggest that intermittent schedules may provide a useful means of testing behavioral impairment in cases where repeated testing is desirable, such as in longitudinal assessment. Learning tasks may be less suitable, unless the types of tasks tested are very different from each other in critical (and perhaps unpredictable) respects. ACKNOWLEDGEMENTS I thank S. Gilbert for programming support, D. Demers, S. Geoffrey, G. Trivett, B. Martin, and V. Liston for technical support, and S. Gilbert and E. Lok for review of the manuscript.

LEAD-INDUCED BEHAVIORAL EFFECTS

245

REFERENCES 1. Bushnell, P. J.; Bowman, R. E. Reversal learning deficits in 22. Petit, T. L.; Le Boutillier, J. C. Effects of lead exposure during young monkeys exposed to lead. Pharmacol. Biochem. Behav. development on neocortical dendritic and synaptic structure. Exp. 10:733-742; 1979. Neurol. 64:482-492; 1979. 2. Bushnell, P. J.; Bowman, R. E. Persistence of impaired reversal 23. Reuhl, K. R.; Rice, D. C.; Gilbert, S. G.; Mallet, J. Effects of learning in young monkeys exposed to low levels of dietary lead. chronic developmental lead exposure on monkey neuroanatomy: J. Toxicol. Envir. Health 5:1015-1023; 1979. visual system. Toxicol. AppL Pharmacol. 99:501-509; 1989. 3. Cory-Slechta, D. A. (1984). The behavioral toxicity of lead: Prob24. Rice, D. C. Operant conditioning of infant monkeys (Macaca lems and perspectives. In: Thompson, R.; Dews, P. B., eds. Adv. fascicularia) for toxicity testing. Nenrobeh. Toxicol. (suppl. 1), Behav. Pharmacol. New York: Academic Press; 1984:211-215. 1:85-92; 1979. 4. Cory-Slechta, D. A.; Pokora, M. J.; Widzowski, D. V. Behav25. Rice, D. C. Behavioral deficit (delayed matching to sample) in ioral manifestations of prolonged lead exposure initiated at difmonkeys exposed from birth to low levels of lead. Toxicol. Appl. ferent stages of the life cycle: II. Delayed spatial alternation. Pharmacol. 75:337-345; 1984. Nenorotoxicol. 12:761-776; 1991. 26. Rice, D. C. Chronic low-lead exposure from birth produces defi5. Cowey, A. (1968). Discrimination. In: L. Weiskrantz, ed. Analycits in discrimination reversal in monkeys. Toxicol. Appl. Pharsis of behavioral change. New York: Harper and Row; 1968:189macol. 77:201-210; 1985. 236. 27. Rice, D. C. Schedule-controlled behavior in infant and juvenile 6. Edgington, E. Randomization tests. New York: Dekker; 1981: monkeys exposed to lead from birth. Neurotoxicol. 9:75-88; 74-85. 1988. 7. Gilbert, S. G.; Rice, D. C. Nova Sked II: A behavioral notation 28. Rice, D. C. Lead-induced behavioral impairment on a spatial language utilizing the Data General Corporation real-time disk discrimination reversal task in monkeys exposed during different operating system. Behav. Research Meth. Instrumen. 11:71-73; periods of development. Toxicni. Appl. Pharmacol. 106:327-333; 1979. 1990. 8. Gilbert, S. G.; Rice, D. C. Low-level life-time lead exposure pro29. Rice, D. C. Behavioral impairment produced by developmental duces behavioral toxicity (spatial discrimination reversal) in adult lead exposure: Evidence from primate research. In: H. L. Needlemonkeys. Toxicol. Appl. Pharmacol. 91:484-490; 1987. man, ed. Human lead exposure, Boca Raton, FL: 1992; 137-154. 9. Goldman, P. S.; Lewis, M. E. Developmental biology of brain 30. Rice, D. C.; Gilbert, S. G. Early chronic low-level methylmercury damage and experience, In: C. W. Cotman, ed. Neuronal plasticpoisoning in monkeys impairs spatial vision. Sci. 216:759-761; ity. New York: Raven Press; 291-310; 1978. 1982. 10. Greenongh, W. T.; Volkmar, F. R. Pattern of dendritic 31. Rice, D. C.; Gilbert, S, G. Low level lead exposure from birth branching in occipital cortex of rats reared in complex environproduces behavioral toxicity (DRL) in monkeys. Toxicol. Appl. ments. Expl. Neurol. 40:491-504; 1973. Pharmacol. 80:421-426; 1985. 11. Harlow, H. F.; Harlow, M. K.; Schiltz, K. A.; Mohr, D. J. The 32. Rice, D. C.; Gilbert, S. G. Sensitive periods for lead-induced effect of early adverse and enriched environments on the learning behavioral impairment (nonspatial discrimination reversal) in ability of rhesus monkeys. In L. E. Jarrard, ed. Cognitive promonkeys. Toxicol. Appl. Pharmacol. 102:101-109; 1990. cesses of nonhuman primates. New York: Academic Press; 1971: 33. Rice, D. C.; Gilbert, S. G. Lack of sensitive period for lead121-148. induced behavioral impairment on a spatial delayed alternation 12. Juraska, J. M.; Greenough, W. T.; Elliott, C.; Mack, K. J.; task in monkeys. Toxicol. Appl. Pharmacol. 103:364-373; 1990. Berkowitz, R. Plasticity in adult rat visual cortex: An examina34. Rice, D. C.; Gilbert, S. G. Automated behavioral procedures for tion of several cell populations after differential rearing. Behav. infant monkeys. Neurotoxicol. Teratol. 12:429-439; 1990. Nenrol. Biol. 29:157-167; 1980. 35. Rice, D. C.; Gilbert, S. G. Effects of developmental exposure to 13. Levin, E. D.; Bowman, R. E. The effect of pre- or postnatal lead methylmercury on spatial and temporal visual function in monexposure on Hamilton Search Task in monkeys. Neurobehav. keys. Toxicol. Appl. Pharmacol. 102:151-163; 1990. Toxicol. Teratol. 5:391-394; 1983. 36. Rice, D. C.; Gilbert, S. G.; Willes, R. F. Neonatal low-level lead 14. Levin, E. D.; Bowman, R. E. (1986). Comparative sensitivity exposure in monkeys (Macacafascicularis): Locomotor activity, of the Hamilton search task and delayed spatial alternation in schedule-controlled behavior, and the effects of amphetamine. detecting long-term effects of neonatal lead exposure in monkeys. Toxicol. Appl. Pharmacol. 51:503-513; 1979. Neurobehav. Toxicol. Teratol. 8,219-224. 37. Rice, D. C.; Karpinski, K. F. Lifetime low-level lead exposure 15. Levin, E. D.; Bowman, R. E. Long-term effects of chronic postproduces deficits in delayed alternation in adult monkeys. Neurotoxicol. Teratol. 10:207-214; 1988. natal lead exposure on delayed spatial alternation in monkeys. 38. Rice, D. C.; Willes, R. F. Neonatal low-level lead exposure in Neurotoxicol. Teratol. 10:505-510; 1988. 16. Lillenthal, H.; Winneke, G.; Brockhaus, A.; Malik, B. Pre- and monkeys (Macacafascicularis): Effects on two-choice non-spatial postnatal lead-exposure in monkeys: Effects on activity and learnform discrimination. J. Environ. Pathol. Toxicol. 2:1195-1203; ing set formation. Neurobehav. Toxicol. Teratol. 8, 265-272; 1979. 1986. 39. Stell, M.; Riesen, A. Effects of early environment on monkey 17. Lorton, D.; Anderson, W. J. Altered pyramidal cell dendritic cortex. Neuroanatomical changes following somatomotor experidevelopment in the motor cortex of lead intoxicated neonatal rats. ence: effects on Layer III pyramidal cells in monkey cortex. BeA Golgi study. Neurobehav. Toxicol. Teratol. 8, 45-50; 1986. hay. Neurosci. 101:341-346; 1987. 40. Stoilnitz, F. Spatial variables, observing response, and discrimi18. Mahajan, D. S.; Desiraju, T. Alterations of dendritic branching and spine densities of hippocampal CA3 pyramidal neurons innation learning sets. Psychol. Rev. 72:247-261; 1965. 41. Tees, R. C.; Midgley, G.; Nesbit, J. C. The effect of early visual duced by operant conditioning in the phase of brain growth spurt. Experimen. Neurol. 100:1-15; 1988. experience on spatial maze learning in rats. Devel. Psychobiol. 19. Mele, P. C.; Bushnell, P. J.; Bowman, R. E. Prolonged behav14:425-438; 1981. ioral effects of early postnatal lead exposure in rhesus monkeys: 42. Venable, N.; Pinto-Hamuy, T.; Arraztoa, J. A.; Contador, M. f'uted interval responding and interactions with scopolamine and T.; Chellow, A.; Peran, C.; Valenzuela, X. Greater efficacy of preweaning than postweaning environmental enrichment on maze pentobarbital. Neurotox. Teratol. 6:129-135; 1984. 20. Morrison, D. F. Multivariate statistical methods. 2nd ed. New learning in adult rats. Behav. Br. Res. 31:89-92; 1981. 43. Winneke, G.; Brockhaus, A.; Baltissen, R. Neurobehavioral and York: McGraw Hill. 1976:216. 21. Petit, T. L.; Alfano, D. P. Differential experience following desystemic effects of long-term blood lead elevation in rats I. Discrimination learning and open field behavior. Arch. Toxicol. 37: velopmental lead exposure: Effects on brain and behavior. Pharmacol. Biochem. Behav. 11:165-171; 1979. 247-263; 1977.