- Email: [email protected]
Performance of normal and neurological mutant mice on radial arm maze and active avoidance tasks
BEHAVIORAL AND NEURAL BIOLOGY
46, 216-226 (1986)
Performance of Normal and Neurological Mutant Mice on Radial Arm Maze and Active Avoidance Tasks DANIEL GOLDOWITZ AND JULIE KOCH 1
Daniel Baugh Institute of Anatomy, Thomas Jefferson University, 1020 Locust Street, Philadelphia, Pennsylvania 19107 The objectives of the present study are to assess the import of gene-imposed structural alterations on behavioral performance and obtain performance data preliminary to studies of experimental mouse chimera behavior. Reeler, staggerer, and weaver neurological mutant, and control B6C3 and ichthyosis mice were tested on radial arm maze and active avoidance tasks. Weaver mice had incapacitating seizures while performing the radial arm maze task and were, therefore, removed from further testing. Staggerer mice displayed a significant deficit on both tasks compared to control mice. Homozygous reeler mice (rl/rI) also had a significant deficit on the active avoidance task compared to + / r l control mice but not significantly poorer than ichthyosis mice. However, their performance on the radial arm maze task, while initially poor, improved so that they performed the task similar to wild-type controls. Three of the reeler mice reached criterion for solving the radial arm maze task. None of the staggerer mice reached criterion. These data are discussed in terms of the value of using neurologically mutant mice in dissecting structural-functional relationships. It is suggested that the behavior of these mutants might point toward specific components of cerebellar involvement in behavioral acts. © 1986AcademicPress, Inc.
The use of genetically defined strains of Mus musculus in biomedical research has greatly facilitated our understanding of biological phenomena. In particular, such work in the behavioral sciences has allowed correlations to be made between structure and function (Wimer & Wimer, 1982). Previous work in the rodent has demonstrated a relationship in "normal" mice between gene-determined variations in hippocampal structure and the performance on various behavioral tasks (Jaffard & Destrade, 1982; Schwegler & Lipp, 1983; Wimer, Wimer, & Wimer, 1983). Specific damage to the rodent hippocampal formations is known to alter the animal's performance on certain behavioral tasks. The radial arm maze is one Support for this research was provided by a National Institute of Neurological and Communicative Disorders and Stroke Grant 2-R23-NS-18716. The authors acknowledge the secretarial help of Sandy Parsons. Dr. Ray Kesner has been a source of great help from the inception to the completion of these studies. Dr. Hyman Menduke has provided statistical aid. 216 0163-1047/86 $3.00 Copyright @ 1986 by Academic Press, Inc. All fights of reproduction in any form reserved.
BEHAVIOR OF NEUROLOGICAL MUTANT MICE
217
such "hippocampal-specific" task (Becker, Walker, & Olton, 1980). The hippocampal formation has been shown to be selectively involved in rodent performance on active avoidance tasks (for review and discussion, see O'Keefe and Nadel (1978)). Furthermore, it has recently been demonstrated that genetically determined changes in hippocampal circuitry affect rodent performance on a two-way active avoidance task (Schwegler & Lipp, 1983). The purpose of this study was to determine if specific behavioral deficits could be correlated with reproducible structural alterations by exploiting neurological mutant mice and their genetically imposed defects of neuronal structure and circuitry. Our initial focus in these studies is on the reeler mouse and behavioral tasks found to be dependent on intact hippocampal function (i.e., radial arm maze and two-way active avoidance). The homozygous reeler mouse (rl/rl) is particularly intriguing because the structural integrity of the reeler hippocampal formation is affected. There is a defective positioning of neurons (as is characteristic for all cortical structures in this mutant (Goffinet, 1984)) and a decreased number of dentate gyrus granule cells and aberrant afferent lamination (Stanfield & Cowan, 1979) in the reeler. The reeler is ataxic, presumably due to cerebellar cortical dysgenesis. Thus, staggerer and weaver mice, two mutants who are also ataxic but with primary cerebellar involvement, were also tested in the behavioral tasks. Heterozygous reeler ( + / r l ) and B6C ichthyosis (ic/ic) mice were used as control, neurologically normal mice. The ichthyosis mouse was used for its value in subsequent chimera studies (Goldowitz & Mullen, 1982) where we plan to correlate the behavioral phenotype of rl/rl ~ ic/ic chimeras with the complement of genetically mutant and genetically normal neurons that comprise various brain regions.
METHODS
Subjects The mice used in all experiments were bred and housed in our animal facilities. Stocks of mice were originally obtained from Jackson Laboratories (Bar Harbor, Maine). Five groups of female mice were used: (1) B6C3H a/a mice heterozygous for the Hammertoe and reeler alleles (B6C3 Hm rl/+ +) (n = 8 maze; n = 9 avoidance); (2) homozygous reelers (B6C3 Hm rl/Hm r/) (n = 4 maze, n = 6 avoidance); (3) C57B1/6 x BALB/cWt ichthyosis (B6C ic/ic) albino mice (n = 8 maze; n = 7 avoidance); (4) B6C3H a/a homozygous staggerer (sg/sg) mice (n = 4 maze; n = 6 avoidance); and (5) homozygous weaver (wv/wv) mice on a hybrid agouti background for the maze. The heterozygous reeler mouse has no known neural deficits, and in pilot studies, had similar performance levels as littermate B6C3H a/a + +/+ + mice. As far as can be observed the Hammertoe gene does not affect behavior and its use is to provide a
218
GOLDOWlTZ AND KOCH
marker for the presence of one or two doses of the reeler gene. Slightly flexed toes indicate a single dose of the H m and rl genes, while a fisted or hammerhead-shaped configuration indicates a double dose of the H m and rl genes. All mice used were adults between 3 and 9 months of age. Mice were first trained on the radial arm maze task, followed by training on the active avoidance task. Due to the greater variability of avoidance scores and the less time-intensive nature of the avoidance task, additional mice were included for avoidance training. Prior to testing mice were provided with food and water ad libitum. Room lights were kept on a 11 h/13 h lights-on/lights-off schedule. Behavioral Apparatuses
The radial arm maze was constructed of plywood painted with white latex paint. The central piece was a 28-cm-wide octagon with 8 equally spaced arms 40 cm long by 9 cm wide extending from the octagonal platform. A goal well 1 cm deep was placed 2 cm from the end of each arm. The maze was placed level on a stool top I m from the ground. The environment around the maze was kept constant, with pictures on two walls, a sink, the experimenter, and utility equipment surrounding the maze. Ataxic mutant mice were also tested on the apparatus and, therefore, sides were necessary. Siding consisted of clear thin acetate sheets or chicken wire. The active avoidance test apparatus (BRS/LVE, Baltimore, Md.) was 19 cm high, 11.4 cm wide, and 27 cm long. The floor was composed of 0.2-cm diameter metal bars with 0.6-cm spaces between the bars. A thin metal barrier between boxes was 2.3 cm high. The CS was a 2.8-KHz tone and a white " Q " lamp light in the " h o m e " compartment. The US was a 4-mA shock. Ten seconds intervened between CS and onset of US and between trials. The avoidance apparatus was housed inside an environmental chamber. Testing Procedures Radial arm maze. Experimental mice were housed in a group cage with 3-4 mice in each experimental run. The group consisted of 1-2 mice from each genotype. Mice were first accustomed to the test apparatus for 10 min a day for 5 days. The shaping procedure prior to testing consisted of food deprivation on Day 1. On subsequent days measured amounts of lab chow were given to bring mice to 80-85% of original body weight. On Days 2 through 5, mice were shaped by exposure to the radial maze with reward pellets (one-fourth piece of Froot Loops weighing about 16 mg) placed in the goal wells in addition to successive spots on each arm leading to the goal wells. On the fifth day of shaping, reward was placed only in the goal well. Testing started after the fifth day of shaping. The reward pellet for mutant mice was reduced to about 9 mg (about one-sixth of a Froot Loop) to equalize the amount of time
BEHAVIOR OF NEUROLOGICAL MUTANT MICE
219
needed at the end of each arm to consume the reward. A trial consisted of loading each goal well with a reward pellet, placing the mouse in the center of the octagonal platform, and allowing the mouse a maximum of 30 min or enough trials until all eight arms were visited and the pellets found and eaten. Animals were tested at a similar time each day for 5 days, with a 2-day recess between 5-day/5-trial sessions. Mice were trained for 30 trials or until they reached a criterion of 3 consecutive trials with eight correct arms chosen in the first 8 trials. A running record of arm choices (novel or repeat) was kept. At the completion of each trial the mice were supplied Purina lab chow sufficient to keep them at the 80-85% original weight level. Active avoidance. Mice were acclimated to the test box for 5 min each day for 2 successive days. The box was illuminated and no shocks were delivered. Testing followed for 10 days with one block of 10 trials/day. The number of avoidances (unshocked crossover), escapes (shocked crossover), and nondisplacements (no crossover) were automatically recorded for each trial.
Analysis of Data For the radial arm maze task, three measures of performance were compared: (1) the number of trials to reach criterion of three consecutive trials without an error, (2) the total number of choices to collect all eight pieces of food over the first and last five trials, and (3) the number of novel (correct) and repeat (incorrect) selections in the first eight choices over the first and last five trials. These data were analyzed using a single factorial analysis of variance, followed by comparison of groups with the Newman-Keuls procedure (Winer, 1962). The mean total number of selections and numbers of novel selections in the first eight per five-trial block were also analyzed relative to the calculated theoretical means as discussed by Eckerman (1980). These theoretical means were calculated by random simulation with the elimination of immediate repeats and reported as 18.18 (no SD given) and 5.61 (SD = 0.83), respectively. A Z score was obtained and significance levels were adjusted for multiple comparisons with the Bonferroni correction (Gill, 1978). For the active avoidance task the percentage of avoidances, escapes and nondisplacements were tabulated. Data were analyzed using a single factor analysis of variance followed by group comparisons as noted above.
RESULTS
I. Radial Arm Maze During our preliminary trials using an open maze, we found that B6C3 a/a and B6C ic mice readily learned the radial arm maze task. The
220
GOLDOWITZ AND KOCH
mouse strategy and behavior was very similar to that of rats as described by Olton and Samuelson (1976). With neurological mutant, ataxic mice it was necessary to erect sides made of clear acetate sheets or chicken wire so that the mutants were prevented from falling to the floor. However, nonataxic mice in this version of the maze now adopted an egocentric or kinesthetic strategy (see Pico & Davis, 1984) in solving the radial arm maze, that is, picking the adjacent arm from the previously chosen arm in either a clockwise or anticlockwise direction. Twenty-one of the twentyfour animals tested in the present experimental series used an egoccentric strategy for solving each trial of the radial arm maze problem. Of the three mice who applied a spatial strategy to the problem's solution, their scores (number of trials to criterion, total choices to get all food and mean correct choices of first eight) were not unlike those of other mice who solved the maze with the kinesthetic strategy. The number of trials for ic, +/rl, and rl/rl mice to adopt an egocentric strategy (defined as five consecutive choices in a given direction for three consecutive trials) were 3.2, 3.0, and 2.2, respectively. It took 6.5 trials before sg/sg mice reached this egocentric approach. The neurologically mutant mice that were used displayed a similar degree of motor dysfunction while in the maze. Staggerer mice moved from arm to arm generally faster than reelers but not as quickly as control mice. Running times of individual mice were not predictive of success in the maze. Testing of weaver mice was discontinued as they all had generalized seizures during maze trials. We have intermittently observed seizure activity in the weaver colony (primarily in heterozygous mutants), and it has been noted by Seyfried (1982). There was a significant strain effect (F(3, 20) = 8.036, p = .001) for the number of trials to criterion. Analysis of the criterion scores by the Newman-Keuls method showed that nonataxic mice performed at a level that was significantly superior to the neurological mutant mice (p < .05) (Fig. 1). The B6C ic and B6C3 + / r l mice started at a level of performance that was significantly better than chance in terms of the mean correct choices of the first eight (7 and 7.2, respectively, compared to chance of 5.6, p < .01). The total number of choices to get all pieces of food was also far less than would be expected from chance performance (Fig. 2). All of the B6C ic mice and six of the eight B6C3 + / r l mice reached a criterion of three consecutive errorless solutions. The variable behavior within a given group is likely due to the segregating backgrounds on which the colony is maintained, although the behavioral results are largely consistent. The neurological mutant mice had a much poorer performance during the initial trials. Over the initial 5 trials rl and sg mice performed at levels that were not different from chance behavior and significantly
BEHAVIOR OF NEUROLOGICAL MUTANT MICE
221
3O-
T
25 ¸ z 20w I-
T
7" 15"
P_ Io-
T,
5"
B6C iclic
B6C3+Irl
rIlrl
sg/sg
STRAI N
FIG. 1. M e a n n u m b e r of trials to reach a criterion of 3 successive, errorless runs. The 4 strain groups are listed by g e n o t y p e on the abcissa. In the radial a r m maze, the n u m b e r s of tested animals were 8, 8, 4, and 4 for B6C ic/ic, B6C3 + / r l , rl/rl, and sg/sg, respectively. T h e standard errors o f the m e a n are marked atop each bar. All four of the sg/sg mice were tested for 30 trials each without reaching criterion.
poorer than the + / r l and ic nonataxic mice (p < .01) (Fig. 2, 3). Over trials, however, reeler mice showed a clear improvement (Fig. 2, 3). The statistically clear effect between strains comparing scores of the last 5 trials (F(13, 20) = 8,48 and 13.66; for mean correct choices and total choices to solve, respectively, both with p values < .001) was due only to the continuing poor performance of the sg mice. The reelers' performance over the last 5 trials was not different from the nonataxic mice. In fact, 24, 20.
t_)
8 8
~ o )- 4"
B6C ic/ic B6C3+/rl rl/rl STRAIN
sg/sg
FIG. 2. M e a n n u m b e r of total choices utilized to get all 8 pieces of food over the first 5 trials (open bars) a n d the last 5 trials (slashed bars). The horizontal slashed line represents the level of p e r f o r m a n c e if b a s e d u p o n r a n d o m choices.
222
GOLDOWITZ AND KOCH
5
4
3
T mz
B6C it/it B6C3 +/rl rl/rl STRAIN
sg/s_?
FIG. 3. Mean number of incorrect choices in first 8 selections over the first set of 5 trials (open bars) and the last set of 5 trials (slashed bars). three of four rl/rl mice reached criterion. None of the sg/sg mice reached criterion. As a group, sg/sg mice had only one errorless trial out of a total of 120.
H. Active Avoidance During the shuttle box acclimation period most of the mice crossed over the barrier at least once; however, cross-overs during this time were not necessarily related to subsequent performance. The barrier presented no obvious encumbrance to cross-overs by the mutant, ataxic mice. There was a significant strain effect (F(3, 24) = 12.861, p < .001) for the percentage avoidance (Fig. 4). Analysis between strains using the Newman-Keuls method indicated that the B6C3 + / r l performance was significantly better than B6C ic and rl/rl mice (p < .05) and sg/sg mice (p < .01). The performance of rl/rl mice was not statistically different from B6C ic mice. Staggerer mice performed at a significantly poorer level than all other strains (p < .01). The heightened reactivity of B6C ic mice to shock may be one aspect to account for their poorer performance than B6C3 + / r l mice. Reeler mice, on the other hand, displayed more of a freezing reaction to the onset of shock. This behavior may account for their poorer performance and greater percentage of nondisplacements (12% compared to less than 2% for each of the three other strains). The performance of staggerer mice did not seem attributable to either of these more obvious behavioral responses.
DISCUSSION Mice can learn to solve a radial arm maze task (Levy, Kluge & Elsmore, 1983; Pico & Davis, 1984; Reinstein, DeBoissiere, Robinson & Wurtman,
B E H A V I O R O F N E U R O L O G I C A L M U T A N T MICE
223
100
9O
T
8O 7O 6O
T,
i
5O
~ 40 o,,~30
T
20 10
B6C'ic/ic B6C3+/rl r l / r l
s~ I s~
STRAIN
FIO. 4. Mean percentage of avoidance over 10 sessions with 10 trials/session. The number o f tested animals were 7, 9, 6, and 6 for B6C ic, B6C3 + / r l , rt/rl, and sg/sg mice, respectively.
1983; and the present study). Depending upon the circumstances of the maze, mice will adopt either a spatial or egocentric strategy in solving the maze. The anatomical substrata of the egocentric strategy are presently unknown and may differ from the well-documented involvement of the hippocampal formation in the spatial solution of the radial arm maze problem (Becker et al., 1980). We found, however, that both strains of the nonataxic mice initially performed at levels that were significantly better than chance. This may reflect the rodent's tendency towards novelty as demonstrated in spontaneous alternation tasks. The intact hippocampal formation has been found to be important for rodents to demonstrate spontaneous alternation (Douglas, 1967). This was the interpretation of the only other study to look at reeler performance in behavioral paradigms, which found that rl/rl mice did not show spontaneous alternation (Bliss & Errington, 1977). This study, however, is problematic in that the rl/rl mice used were very young and in dubious health, such that the authors noted that testing only continued until postnatal Day 24, at which time "owing to increasing weakness, Reelers could not be tested." Using hybrid rl/rl mice, we were able to avoid this problem. Our present findings, however, do not suggest a correlation between rl hippocampal structural deficits and behavioral performance on the radial arm maze task. This finding can be placed among a growing list of items, indicating that reeler brain function might be more normal than one might predict based upon the gross cortical cell positioning defects and cell loss seen in this mutant (e.g., Caviness & Frost, 1983; Dr~ger, 1981; Simmons & Pearlman, 1982; Steindler & Colwell, 1976).
224
GOLDOWlTZ AND KOCH
In contrast to the radial arm maze test, the performances of rl/rl and B6C ic mice on the active avoidance task were significantly poorer than B6C3 + / r l mice. This behavioral difference between B6C3 + / r l and B6C ic mice may be attributable to documented subtle differences in hippocampal structure (Schwegler & Lipp, 1983; Wimer et al., 1983). If behavioral differences on this task can be attributed to specific variations in hippocampal structure, then the reeler does not offer an easy explanation. The fact that granule cell number and density is reduced in the reeler (Stanfield & Cowan, 1979) does not neatly fit in with previously determined correlations between these variables and avoidance behavior in normal mice (Wimer et al., 1983). On the other hand, the increased occurrence of mossy fiber connections on the basal dendrites of rl regio inferior pyramidal cells (Stanfield & Cowan, 1979) is concordant with the negative correlation found between such contacts and avoidance behavior (Schwegler & Lipp, 1983). This point, of course, needs to be balanced with the even poorer performance of sg mice and their evidently normal hippocampal formation (unpublished observations). Indeed, the poor performance of staggerer (sg) mutant mice was a surprise. This mutant strain was chosen as the "ataxic control" strain for reeler, that is, a mouse with gross cerebellar dysgenesis and ataxia but lacking the neocortical and hippocampal abnormalities seen in the reeler mutant. The poor performance of sg mice on the radial arm maze and active avoidance tasks does not appear explicable by greater deficits in motor behavior or due to genetic background. Staggerer mice had shorter running times between goal boxes than did reeler mice in the radial arm maze and demonstrated a better escape activity in the shuttle box. Both mutant strains were on the B6C3 a/a Fl-hybrid background. Subsequent matings in our colony have been among brothers and sisters. This has created varying genetic lines within each colony that should be equally dissimilar among individual reelers as well as staggerers. Instead, it would appear that the observed performance differences between these mutants would involve the dissection of gene-imposed deficits in staggerer compared to reeler mutants. Such a difference may reside in the cerebellum since this is believed to be the primary site of staggerer mutant gene action (Sidman, Lane & Dickie, 1%2). The major identified morphological abnormalities in staggerer are a failure of granule cell parallel fibers to make contact with Purkinje cells, granule cell death and fewer Purkinje cells, and Purkinje cell dysgenesis (Landis & Sidman, 1978; Sidman, 1968). The reeler cerebellar abnormality resides in a greatly disturbed neuronal migration resulting in many improperly positioned Purkinje cells although proper connectivity between granule and Purkinje cells is preserved in the small population of correctly positioned Purkinje cells (Mariani, Crepel, Mikoshiba, Changeux, & Sotelo, 1977). It is interesting to note in this regard that rats with experimentally induced agranular
BEHAVIOR OF NEUROLOGICAL MUTANT MICE
225
cerebella, which effectively eliminates granule cells and the parallel fiberPurkinje cell contact, fail to show spontaneous alternation behavior (Pellegrino & Altman, 1979). These considerations point to a possible dissection of a cerebellar component to learned solutions of behavioral problems. It is thought that the cerebellum is importantly involved in learned behavior (see Ito, 1984; McCormick, Clark, Lavond, & Thompson, 1982). More direct tests of cerebellar function would seem to be appropriate relative to the present findings. Indeed, neurological mouse mutants offer a unique and relatively unexploited means to explore such structural-functional correlations. REFERENCES Bliss, T. V. P., & Errington, M. L. (1977). 'Reeler' mutant mice fail to show spontaneous alternation. Brain Research, 124, 168-170. Becker, J. T., Walker, J. A., & Olton, D. S. (1980). Neuroanatomical bases of spatial memory. Brain Research, 200, 307-320. Caviness, V. S., Jr., & Frost, D. O. (1983). Thalamocortical projections in the reeler mutant mouse. Journal of Comparative Neurology, 219, 182-202. Douglas, R. J. (1967). The hippocampus and behavior. Psychological Bulletin, 67, 416442. Drfiger, U. C. (1981). Observations on the organization of the visual cortex in the reeler mouse. Journal of Comparative Neurology, 201, 555-570. Eckerman, D. A. (1980). Monte Carlo estimation of chance performance for the radial arm maze. Bulletin of the Psychonomic Society, 15, 93-95. Gill, J. L. (1978). Design and analysis of experiments in the animal and medical sciences (pp. 175-177). Ames, Iowa: Iowa State Univ. Press. Goffinet, A. M. (1984). Events governing organization of postmigratory neurons: Studies on brain development in normal and reeler mice. Brain Research Reviews, 7, 261296. Goldowitz, D., & Mullen, R. J. (1982). Use of the mouse mutant ichthyosis to study cellcell interactions in brain. Developmental Biology, 89, 261-267. Ito, M. (1984). The cerebellum and neural control (pp. 436-451). New York: Raven Press. Jaffard, R., & Destrade, C. (1982). Learning and memory processes as related to genotypic or experimental variations of hippocampal cholinergic activity in inbred strains of mice. In I. Leiblich (Ed.), Genetics of the brain (pp. 300-322). Amsterdam: Elsevier. Landis, D. M. D., & Sidman, R. L. (1978). Electron microscopic analysis of postnatal histogenesis in the cerebellar cortex of staggerer mutant mice. Journal of Comparative Neurology, 179, 831-864. Levy, A., Kluge, P. B., & Elsmore, T. F. (1983). Radial arm maze performance of mice: Acquisition and atropine effects. Behavioral and Neural Biology, 39, 229-240. Mariani, J., Crepel, F., Mikoshiba, K., Changeux, J. P., & Sotelo, C. (1977). Anatomical, physiological and biochemical studies of the cerebellum from reeler mutant mouse. Philosophical Transactions of the Royal Society of London, 281, 1-28. McCormick, D. A., Clark, G. A., Lavond, D. G., & Thompson, R. F. (1982). Initial localization of the memory trace for a basic form of learning. Proceedings of the National Academy of Sciences of the United States of America 79(8), 2731-2742. O'Keefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map (pp. 308-310; Table A22). Oxford: Clavendon. Olton, D. S., & Samuelson, R. J. (1976). Remembrances of places passed: Spatial memory in rats. Journal of Experimental Psychology: Animal Behavior Processes, 2, 97-116.
226
GOLDOWITZ AND KOCH
Pellegrino, L. J., & Altman, J. (1979). Effects of differential interference with postnatal cerebellar neurogenesis on motor performance, activity level and maze learning of rats: A developmental study. Journal of Comparative Physiology and Psychology 93, 1-33. Pico, R. M., & Davis, J. L. (1984). The radial arm maze performance of mice: Assessing the dimensional requirements for serial order memory in animals. Behavioral and Neural Biology, 40, 5-26. Reinstein, D. K., DeBoissiere, T., Robinson, N., & Wurtman, R. J. (1983). Radial maze performance in three strains of mice: Role of the fimbria/fornix. Brain Research, 263, 172-176. Schwegler, H., & Lipp, H. P. (1983). Hereditary covariations of neuronal circuitry and behavior: Correlations between the proportions of hippocampal synaptic fields in the regio inferior and two-way avoidance in mice and rats. Behavioral Brain Research, 7, 1-38. Seyfried, T. N. (1982). Convulsive disorders. In H. L. Foster, J. D. Small, & J. G. Fox (Eds.), The mouse in biomedical research (Vol. 4, pp. 97-124). New York: Academic Press. Sidman, R. L. (1968). Development of interneuronal connections in brains of mutant mice. In F. D. Carlson (Ed.), Physiological and biochemical aspects of nervous integration (pp. 163-193). Englewood Cliffs, N.J.: Prentice-Hall. Sidman, R. L., Lane, P. W., & Dickie, M. M. (1962). Staggerer: A new mutation in the mouse affecting the cerebellum. Science 137, 610-612. Simmons, P. A., & Pearlman, A. L. (1982). Retinopic organization of striate cortex (area 17) in the reeler mutant mouse. Developmental Brain Research, 4, 124-126. Stanfield, B. B., & Cowan, W. M. (1979). The morphology of the hippocampus and dentate gyrus in normal and reeler mice. Journal of Comparative Neurology, 185, 393-422. Steindler, D. A., and Colwell, S. A. (1976). Reeler mutant mouse: Maintenance of appropriate and reciprocal connections in the cerebral cortex and thalamus. Brain Research, 105, 386-393. Wimer, C., Wimer, R. E., & Wimer, J. S. (1983). An association between granule cell density in the dentate gyrus and two-way avoidance conditioning in the house mouse. Behavioral Neuroscience, 97, 844-856. Wimer, R. E., & Wimer, C. C. (1982). A geneticist's map of the mouse brain. In I. Lieblich (Ed.), Genetics of" the brain (pp. 396-420). Amsterdam: Elsevier. Winer, B. J. (1962). Statistical principles in experimental design (pp. 77-85). New York: McGraw-Hill.
46, 216-226 (1986)
Performance of Normal and Neurological Mutant Mice on Radial Arm Maze and Active Avoidance Tasks DANIEL GOLDOWITZ AND JULIE KOCH 1
Daniel Baugh Institute of Anatomy, Thomas Jefferson University, 1020 Locust Street, Philadelphia, Pennsylvania 19107 The objectives of the present study are to assess the import of gene-imposed structural alterations on behavioral performance and obtain performance data preliminary to studies of experimental mouse chimera behavior. Reeler, staggerer, and weaver neurological mutant, and control B6C3 and ichthyosis mice were tested on radial arm maze and active avoidance tasks. Weaver mice had incapacitating seizures while performing the radial arm maze task and were, therefore, removed from further testing. Staggerer mice displayed a significant deficit on both tasks compared to control mice. Homozygous reeler mice (rl/rI) also had a significant deficit on the active avoidance task compared to + / r l control mice but not significantly poorer than ichthyosis mice. However, their performance on the radial arm maze task, while initially poor, improved so that they performed the task similar to wild-type controls. Three of the reeler mice reached criterion for solving the radial arm maze task. None of the staggerer mice reached criterion. These data are discussed in terms of the value of using neurologically mutant mice in dissecting structural-functional relationships. It is suggested that the behavior of these mutants might point toward specific components of cerebellar involvement in behavioral acts. © 1986AcademicPress, Inc.
The use of genetically defined strains of Mus musculus in biomedical research has greatly facilitated our understanding of biological phenomena. In particular, such work in the behavioral sciences has allowed correlations to be made between structure and function (Wimer & Wimer, 1982). Previous work in the rodent has demonstrated a relationship in "normal" mice between gene-determined variations in hippocampal structure and the performance on various behavioral tasks (Jaffard & Destrade, 1982; Schwegler & Lipp, 1983; Wimer, Wimer, & Wimer, 1983). Specific damage to the rodent hippocampal formations is known to alter the animal's performance on certain behavioral tasks. The radial arm maze is one Support for this research was provided by a National Institute of Neurological and Communicative Disorders and Stroke Grant 2-R23-NS-18716. The authors acknowledge the secretarial help of Sandy Parsons. Dr. Ray Kesner has been a source of great help from the inception to the completion of these studies. Dr. Hyman Menduke has provided statistical aid. 216 0163-1047/86 $3.00 Copyright @ 1986 by Academic Press, Inc. All fights of reproduction in any form reserved.
BEHAVIOR OF NEUROLOGICAL MUTANT MICE
217
such "hippocampal-specific" task (Becker, Walker, & Olton, 1980). The hippocampal formation has been shown to be selectively involved in rodent performance on active avoidance tasks (for review and discussion, see O'Keefe and Nadel (1978)). Furthermore, it has recently been demonstrated that genetically determined changes in hippocampal circuitry affect rodent performance on a two-way active avoidance task (Schwegler & Lipp, 1983). The purpose of this study was to determine if specific behavioral deficits could be correlated with reproducible structural alterations by exploiting neurological mutant mice and their genetically imposed defects of neuronal structure and circuitry. Our initial focus in these studies is on the reeler mouse and behavioral tasks found to be dependent on intact hippocampal function (i.e., radial arm maze and two-way active avoidance). The homozygous reeler mouse (rl/rl) is particularly intriguing because the structural integrity of the reeler hippocampal formation is affected. There is a defective positioning of neurons (as is characteristic for all cortical structures in this mutant (Goffinet, 1984)) and a decreased number of dentate gyrus granule cells and aberrant afferent lamination (Stanfield & Cowan, 1979) in the reeler. The reeler is ataxic, presumably due to cerebellar cortical dysgenesis. Thus, staggerer and weaver mice, two mutants who are also ataxic but with primary cerebellar involvement, were also tested in the behavioral tasks. Heterozygous reeler ( + / r l ) and B6C ichthyosis (ic/ic) mice were used as control, neurologically normal mice. The ichthyosis mouse was used for its value in subsequent chimera studies (Goldowitz & Mullen, 1982) where we plan to correlate the behavioral phenotype of rl/rl ~ ic/ic chimeras with the complement of genetically mutant and genetically normal neurons that comprise various brain regions.
METHODS
Subjects The mice used in all experiments were bred and housed in our animal facilities. Stocks of mice were originally obtained from Jackson Laboratories (Bar Harbor, Maine). Five groups of female mice were used: (1) B6C3H a/a mice heterozygous for the Hammertoe and reeler alleles (B6C3 Hm rl/+ +) (n = 8 maze; n = 9 avoidance); (2) homozygous reelers (B6C3 Hm rl/Hm r/) (n = 4 maze, n = 6 avoidance); (3) C57B1/6 x BALB/cWt ichthyosis (B6C ic/ic) albino mice (n = 8 maze; n = 7 avoidance); (4) B6C3H a/a homozygous staggerer (sg/sg) mice (n = 4 maze; n = 6 avoidance); and (5) homozygous weaver (wv/wv) mice on a hybrid agouti background for the maze. The heterozygous reeler mouse has no known neural deficits, and in pilot studies, had similar performance levels as littermate B6C3H a/a + +/+ + mice. As far as can be observed the Hammertoe gene does not affect behavior and its use is to provide a
218
GOLDOWlTZ AND KOCH
marker for the presence of one or two doses of the reeler gene. Slightly flexed toes indicate a single dose of the H m and rl genes, while a fisted or hammerhead-shaped configuration indicates a double dose of the H m and rl genes. All mice used were adults between 3 and 9 months of age. Mice were first trained on the radial arm maze task, followed by training on the active avoidance task. Due to the greater variability of avoidance scores and the less time-intensive nature of the avoidance task, additional mice were included for avoidance training. Prior to testing mice were provided with food and water ad libitum. Room lights were kept on a 11 h/13 h lights-on/lights-off schedule. Behavioral Apparatuses
The radial arm maze was constructed of plywood painted with white latex paint. The central piece was a 28-cm-wide octagon with 8 equally spaced arms 40 cm long by 9 cm wide extending from the octagonal platform. A goal well 1 cm deep was placed 2 cm from the end of each arm. The maze was placed level on a stool top I m from the ground. The environment around the maze was kept constant, with pictures on two walls, a sink, the experimenter, and utility equipment surrounding the maze. Ataxic mutant mice were also tested on the apparatus and, therefore, sides were necessary. Siding consisted of clear thin acetate sheets or chicken wire. The active avoidance test apparatus (BRS/LVE, Baltimore, Md.) was 19 cm high, 11.4 cm wide, and 27 cm long. The floor was composed of 0.2-cm diameter metal bars with 0.6-cm spaces between the bars. A thin metal barrier between boxes was 2.3 cm high. The CS was a 2.8-KHz tone and a white " Q " lamp light in the " h o m e " compartment. The US was a 4-mA shock. Ten seconds intervened between CS and onset of US and between trials. The avoidance apparatus was housed inside an environmental chamber. Testing Procedures Radial arm maze. Experimental mice were housed in a group cage with 3-4 mice in each experimental run. The group consisted of 1-2 mice from each genotype. Mice were first accustomed to the test apparatus for 10 min a day for 5 days. The shaping procedure prior to testing consisted of food deprivation on Day 1. On subsequent days measured amounts of lab chow were given to bring mice to 80-85% of original body weight. On Days 2 through 5, mice were shaped by exposure to the radial maze with reward pellets (one-fourth piece of Froot Loops weighing about 16 mg) placed in the goal wells in addition to successive spots on each arm leading to the goal wells. On the fifth day of shaping, reward was placed only in the goal well. Testing started after the fifth day of shaping. The reward pellet for mutant mice was reduced to about 9 mg (about one-sixth of a Froot Loop) to equalize the amount of time
BEHAVIOR OF NEUROLOGICAL MUTANT MICE
219
needed at the end of each arm to consume the reward. A trial consisted of loading each goal well with a reward pellet, placing the mouse in the center of the octagonal platform, and allowing the mouse a maximum of 30 min or enough trials until all eight arms were visited and the pellets found and eaten. Animals were tested at a similar time each day for 5 days, with a 2-day recess between 5-day/5-trial sessions. Mice were trained for 30 trials or until they reached a criterion of 3 consecutive trials with eight correct arms chosen in the first 8 trials. A running record of arm choices (novel or repeat) was kept. At the completion of each trial the mice were supplied Purina lab chow sufficient to keep them at the 80-85% original weight level. Active avoidance. Mice were acclimated to the test box for 5 min each day for 2 successive days. The box was illuminated and no shocks were delivered. Testing followed for 10 days with one block of 10 trials/day. The number of avoidances (unshocked crossover), escapes (shocked crossover), and nondisplacements (no crossover) were automatically recorded for each trial.
Analysis of Data For the radial arm maze task, three measures of performance were compared: (1) the number of trials to reach criterion of three consecutive trials without an error, (2) the total number of choices to collect all eight pieces of food over the first and last five trials, and (3) the number of novel (correct) and repeat (incorrect) selections in the first eight choices over the first and last five trials. These data were analyzed using a single factorial analysis of variance, followed by comparison of groups with the Newman-Keuls procedure (Winer, 1962). The mean total number of selections and numbers of novel selections in the first eight per five-trial block were also analyzed relative to the calculated theoretical means as discussed by Eckerman (1980). These theoretical means were calculated by random simulation with the elimination of immediate repeats and reported as 18.18 (no SD given) and 5.61 (SD = 0.83), respectively. A Z score was obtained and significance levels were adjusted for multiple comparisons with the Bonferroni correction (Gill, 1978). For the active avoidance task the percentage of avoidances, escapes and nondisplacements were tabulated. Data were analyzed using a single factor analysis of variance followed by group comparisons as noted above.
RESULTS
I. Radial Arm Maze During our preliminary trials using an open maze, we found that B6C3 a/a and B6C ic mice readily learned the radial arm maze task. The
220
GOLDOWITZ AND KOCH
mouse strategy and behavior was very similar to that of rats as described by Olton and Samuelson (1976). With neurological mutant, ataxic mice it was necessary to erect sides made of clear acetate sheets or chicken wire so that the mutants were prevented from falling to the floor. However, nonataxic mice in this version of the maze now adopted an egocentric or kinesthetic strategy (see Pico & Davis, 1984) in solving the radial arm maze, that is, picking the adjacent arm from the previously chosen arm in either a clockwise or anticlockwise direction. Twenty-one of the twentyfour animals tested in the present experimental series used an egoccentric strategy for solving each trial of the radial arm maze problem. Of the three mice who applied a spatial strategy to the problem's solution, their scores (number of trials to criterion, total choices to get all food and mean correct choices of first eight) were not unlike those of other mice who solved the maze with the kinesthetic strategy. The number of trials for ic, +/rl, and rl/rl mice to adopt an egocentric strategy (defined as five consecutive choices in a given direction for three consecutive trials) were 3.2, 3.0, and 2.2, respectively. It took 6.5 trials before sg/sg mice reached this egocentric approach. The neurologically mutant mice that were used displayed a similar degree of motor dysfunction while in the maze. Staggerer mice moved from arm to arm generally faster than reelers but not as quickly as control mice. Running times of individual mice were not predictive of success in the maze. Testing of weaver mice was discontinued as they all had generalized seizures during maze trials. We have intermittently observed seizure activity in the weaver colony (primarily in heterozygous mutants), and it has been noted by Seyfried (1982). There was a significant strain effect (F(3, 20) = 8.036, p = .001) for the number of trials to criterion. Analysis of the criterion scores by the Newman-Keuls method showed that nonataxic mice performed at a level that was significantly superior to the neurological mutant mice (p < .05) (Fig. 1). The B6C ic and B6C3 + / r l mice started at a level of performance that was significantly better than chance in terms of the mean correct choices of the first eight (7 and 7.2, respectively, compared to chance of 5.6, p < .01). The total number of choices to get all pieces of food was also far less than would be expected from chance performance (Fig. 2). All of the B6C ic mice and six of the eight B6C3 + / r l mice reached a criterion of three consecutive errorless solutions. The variable behavior within a given group is likely due to the segregating backgrounds on which the colony is maintained, although the behavioral results are largely consistent. The neurological mutant mice had a much poorer performance during the initial trials. Over the initial 5 trials rl and sg mice performed at levels that were not different from chance behavior and significantly
BEHAVIOR OF NEUROLOGICAL MUTANT MICE
221
3O-
T
25 ¸ z 20w I-
T
7" 15"
P_ Io-
T,
5"
B6C iclic
B6C3+Irl
rIlrl
sg/sg
STRAI N
FIG. 1. M e a n n u m b e r of trials to reach a criterion of 3 successive, errorless runs. The 4 strain groups are listed by g e n o t y p e on the abcissa. In the radial a r m maze, the n u m b e r s of tested animals were 8, 8, 4, and 4 for B6C ic/ic, B6C3 + / r l , rl/rl, and sg/sg, respectively. T h e standard errors o f the m e a n are marked atop each bar. All four of the sg/sg mice were tested for 30 trials each without reaching criterion.
poorer than the + / r l and ic nonataxic mice (p < .01) (Fig. 2, 3). Over trials, however, reeler mice showed a clear improvement (Fig. 2, 3). The statistically clear effect between strains comparing scores of the last 5 trials (F(13, 20) = 8,48 and 13.66; for mean correct choices and total choices to solve, respectively, both with p values < .001) was due only to the continuing poor performance of the sg mice. The reelers' performance over the last 5 trials was not different from the nonataxic mice. In fact, 24, 20.
t_)
8 8
~ o )- 4"
B6C ic/ic B6C3+/rl rl/rl STRAIN
sg/sg
FIG. 2. M e a n n u m b e r of total choices utilized to get all 8 pieces of food over the first 5 trials (open bars) a n d the last 5 trials (slashed bars). The horizontal slashed line represents the level of p e r f o r m a n c e if b a s e d u p o n r a n d o m choices.
222
GOLDOWITZ AND KOCH
5
4
3
T mz
B6C it/it B6C3 +/rl rl/rl STRAIN
sg/s_?
FIG. 3. Mean number of incorrect choices in first 8 selections over the first set of 5 trials (open bars) and the last set of 5 trials (slashed bars). three of four rl/rl mice reached criterion. None of the sg/sg mice reached criterion. As a group, sg/sg mice had only one errorless trial out of a total of 120.
H. Active Avoidance During the shuttle box acclimation period most of the mice crossed over the barrier at least once; however, cross-overs during this time were not necessarily related to subsequent performance. The barrier presented no obvious encumbrance to cross-overs by the mutant, ataxic mice. There was a significant strain effect (F(3, 24) = 12.861, p < .001) for the percentage avoidance (Fig. 4). Analysis between strains using the Newman-Keuls method indicated that the B6C3 + / r l performance was significantly better than B6C ic and rl/rl mice (p < .05) and sg/sg mice (p < .01). The performance of rl/rl mice was not statistically different from B6C ic mice. Staggerer mice performed at a significantly poorer level than all other strains (p < .01). The heightened reactivity of B6C ic mice to shock may be one aspect to account for their poorer performance than B6C3 + / r l mice. Reeler mice, on the other hand, displayed more of a freezing reaction to the onset of shock. This behavior may account for their poorer performance and greater percentage of nondisplacements (12% compared to less than 2% for each of the three other strains). The performance of staggerer mice did not seem attributable to either of these more obvious behavioral responses.
DISCUSSION Mice can learn to solve a radial arm maze task (Levy, Kluge & Elsmore, 1983; Pico & Davis, 1984; Reinstein, DeBoissiere, Robinson & Wurtman,
B E H A V I O R O F N E U R O L O G I C A L M U T A N T MICE
223
100
9O
T
8O 7O 6O
T,
i
5O
~ 40 o,,~30
T
20 10
B6C'ic/ic B6C3+/rl r l / r l
s~ I s~
STRAIN
FIO. 4. Mean percentage of avoidance over 10 sessions with 10 trials/session. The number o f tested animals were 7, 9, 6, and 6 for B6C ic, B6C3 + / r l , rt/rl, and sg/sg mice, respectively.
1983; and the present study). Depending upon the circumstances of the maze, mice will adopt either a spatial or egocentric strategy in solving the maze. The anatomical substrata of the egocentric strategy are presently unknown and may differ from the well-documented involvement of the hippocampal formation in the spatial solution of the radial arm maze problem (Becker et al., 1980). We found, however, that both strains of the nonataxic mice initially performed at levels that were significantly better than chance. This may reflect the rodent's tendency towards novelty as demonstrated in spontaneous alternation tasks. The intact hippocampal formation has been found to be important for rodents to demonstrate spontaneous alternation (Douglas, 1967). This was the interpretation of the only other study to look at reeler performance in behavioral paradigms, which found that rl/rl mice did not show spontaneous alternation (Bliss & Errington, 1977). This study, however, is problematic in that the rl/rl mice used were very young and in dubious health, such that the authors noted that testing only continued until postnatal Day 24, at which time "owing to increasing weakness, Reelers could not be tested." Using hybrid rl/rl mice, we were able to avoid this problem. Our present findings, however, do not suggest a correlation between rl hippocampal structural deficits and behavioral performance on the radial arm maze task. This finding can be placed among a growing list of items, indicating that reeler brain function might be more normal than one might predict based upon the gross cortical cell positioning defects and cell loss seen in this mutant (e.g., Caviness & Frost, 1983; Dr~ger, 1981; Simmons & Pearlman, 1982; Steindler & Colwell, 1976).
224
GOLDOWlTZ AND KOCH
In contrast to the radial arm maze test, the performances of rl/rl and B6C ic mice on the active avoidance task were significantly poorer than B6C3 + / r l mice. This behavioral difference between B6C3 + / r l and B6C ic mice may be attributable to documented subtle differences in hippocampal structure (Schwegler & Lipp, 1983; Wimer et al., 1983). If behavioral differences on this task can be attributed to specific variations in hippocampal structure, then the reeler does not offer an easy explanation. The fact that granule cell number and density is reduced in the reeler (Stanfield & Cowan, 1979) does not neatly fit in with previously determined correlations between these variables and avoidance behavior in normal mice (Wimer et al., 1983). On the other hand, the increased occurrence of mossy fiber connections on the basal dendrites of rl regio inferior pyramidal cells (Stanfield & Cowan, 1979) is concordant with the negative correlation found between such contacts and avoidance behavior (Schwegler & Lipp, 1983). This point, of course, needs to be balanced with the even poorer performance of sg mice and their evidently normal hippocampal formation (unpublished observations). Indeed, the poor performance of staggerer (sg) mutant mice was a surprise. This mutant strain was chosen as the "ataxic control" strain for reeler, that is, a mouse with gross cerebellar dysgenesis and ataxia but lacking the neocortical and hippocampal abnormalities seen in the reeler mutant. The poor performance of sg mice on the radial arm maze and active avoidance tasks does not appear explicable by greater deficits in motor behavior or due to genetic background. Staggerer mice had shorter running times between goal boxes than did reeler mice in the radial arm maze and demonstrated a better escape activity in the shuttle box. Both mutant strains were on the B6C3 a/a Fl-hybrid background. Subsequent matings in our colony have been among brothers and sisters. This has created varying genetic lines within each colony that should be equally dissimilar among individual reelers as well as staggerers. Instead, it would appear that the observed performance differences between these mutants would involve the dissection of gene-imposed deficits in staggerer compared to reeler mutants. Such a difference may reside in the cerebellum since this is believed to be the primary site of staggerer mutant gene action (Sidman, Lane & Dickie, 1%2). The major identified morphological abnormalities in staggerer are a failure of granule cell parallel fibers to make contact with Purkinje cells, granule cell death and fewer Purkinje cells, and Purkinje cell dysgenesis (Landis & Sidman, 1978; Sidman, 1968). The reeler cerebellar abnormality resides in a greatly disturbed neuronal migration resulting in many improperly positioned Purkinje cells although proper connectivity between granule and Purkinje cells is preserved in the small population of correctly positioned Purkinje cells (Mariani, Crepel, Mikoshiba, Changeux, & Sotelo, 1977). It is interesting to note in this regard that rats with experimentally induced agranular
BEHAVIOR OF NEUROLOGICAL MUTANT MICE
225
cerebella, which effectively eliminates granule cells and the parallel fiberPurkinje cell contact, fail to show spontaneous alternation behavior (Pellegrino & Altman, 1979). These considerations point to a possible dissection of a cerebellar component to learned solutions of behavioral problems. It is thought that the cerebellum is importantly involved in learned behavior (see Ito, 1984; McCormick, Clark, Lavond, & Thompson, 1982). More direct tests of cerebellar function would seem to be appropriate relative to the present findings. Indeed, neurological mouse mutants offer a unique and relatively unexploited means to explore such structural-functional correlations. REFERENCES Bliss, T. V. P., & Errington, M. L. (1977). 'Reeler' mutant mice fail to show spontaneous alternation. Brain Research, 124, 168-170. Becker, J. T., Walker, J. A., & Olton, D. S. (1980). Neuroanatomical bases of spatial memory. Brain Research, 200, 307-320. Caviness, V. S., Jr., & Frost, D. O. (1983). Thalamocortical projections in the reeler mutant mouse. Journal of Comparative Neurology, 219, 182-202. Douglas, R. J. (1967). The hippocampus and behavior. Psychological Bulletin, 67, 416442. Drfiger, U. C. (1981). Observations on the organization of the visual cortex in the reeler mouse. Journal of Comparative Neurology, 201, 555-570. Eckerman, D. A. (1980). Monte Carlo estimation of chance performance for the radial arm maze. Bulletin of the Psychonomic Society, 15, 93-95. Gill, J. L. (1978). Design and analysis of experiments in the animal and medical sciences (pp. 175-177). Ames, Iowa: Iowa State Univ. Press. Goffinet, A. M. (1984). Events governing organization of postmigratory neurons: Studies on brain development in normal and reeler mice. Brain Research Reviews, 7, 261296. Goldowitz, D., & Mullen, R. J. (1982). Use of the mouse mutant ichthyosis to study cellcell interactions in brain. Developmental Biology, 89, 261-267. Ito, M. (1984). The cerebellum and neural control (pp. 436-451). New York: Raven Press. Jaffard, R., & Destrade, C. (1982). Learning and memory processes as related to genotypic or experimental variations of hippocampal cholinergic activity in inbred strains of mice. In I. Leiblich (Ed.), Genetics of the brain (pp. 300-322). Amsterdam: Elsevier. Landis, D. M. D., & Sidman, R. L. (1978). Electron microscopic analysis of postnatal histogenesis in the cerebellar cortex of staggerer mutant mice. Journal of Comparative Neurology, 179, 831-864. Levy, A., Kluge, P. B., & Elsmore, T. F. (1983). Radial arm maze performance of mice: Acquisition and atropine effects. Behavioral and Neural Biology, 39, 229-240. Mariani, J., Crepel, F., Mikoshiba, K., Changeux, J. P., & Sotelo, C. (1977). Anatomical, physiological and biochemical studies of the cerebellum from reeler mutant mouse. Philosophical Transactions of the Royal Society of London, 281, 1-28. McCormick, D. A., Clark, G. A., Lavond, D. G., & Thompson, R. F. (1982). Initial localization of the memory trace for a basic form of learning. Proceedings of the National Academy of Sciences of the United States of America 79(8), 2731-2742. O'Keefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map (pp. 308-310; Table A22). Oxford: Clavendon. Olton, D. S., & Samuelson, R. J. (1976). Remembrances of places passed: Spatial memory in rats. Journal of Experimental Psychology: Animal Behavior Processes, 2, 97-116.
226
GOLDOWITZ AND KOCH
Pellegrino, L. J., & Altman, J. (1979). Effects of differential interference with postnatal cerebellar neurogenesis on motor performance, activity level and maze learning of rats: A developmental study. Journal of Comparative Physiology and Psychology 93, 1-33. Pico, R. M., & Davis, J. L. (1984). The radial arm maze performance of mice: Assessing the dimensional requirements for serial order memory in animals. Behavioral and Neural Biology, 40, 5-26. Reinstein, D. K., DeBoissiere, T., Robinson, N., & Wurtman, R. J. (1983). Radial maze performance in three strains of mice: Role of the fimbria/fornix. Brain Research, 263, 172-176. Schwegler, H., & Lipp, H. P. (1983). Hereditary covariations of neuronal circuitry and behavior: Correlations between the proportions of hippocampal synaptic fields in the regio inferior and two-way avoidance in mice and rats. Behavioral Brain Research, 7, 1-38. Seyfried, T. N. (1982). Convulsive disorders. In H. L. Foster, J. D. Small, & J. G. Fox (Eds.), The mouse in biomedical research (Vol. 4, pp. 97-124). New York: Academic Press. Sidman, R. L. (1968). Development of interneuronal connections in brains of mutant mice. In F. D. Carlson (Ed.), Physiological and biochemical aspects of nervous integration (pp. 163-193). Englewood Cliffs, N.J.: Prentice-Hall. Sidman, R. L., Lane, P. W., & Dickie, M. M. (1962). Staggerer: A new mutation in the mouse affecting the cerebellum. Science 137, 610-612. Simmons, P. A., & Pearlman, A. L. (1982). Retinopic organization of striate cortex (area 17) in the reeler mutant mouse. Developmental Brain Research, 4, 124-126. Stanfield, B. B., & Cowan, W. M. (1979). The morphology of the hippocampus and dentate gyrus in normal and reeler mice. Journal of Comparative Neurology, 185, 393-422. Steindler, D. A., and Colwell, S. A. (1976). Reeler mutant mouse: Maintenance of appropriate and reciprocal connections in the cerebral cortex and thalamus. Brain Research, 105, 386-393. Wimer, C., Wimer, R. E., & Wimer, J. S. (1983). An association between granule cell density in the dentate gyrus and two-way avoidance conditioning in the house mouse. Behavioral Neuroscience, 97, 844-856. Wimer, R. E., & Wimer, C. C. (1982). A geneticist's map of the mouse brain. In I. Lieblich (Ed.), Genetics of" the brain (pp. 396-420). Amsterdam: Elsevier. Winer, B. J. (1962). Statistical principles in experimental design (pp. 77-85). New York: McGraw-Hill.