ACTA NEUROBIOL. EXP. 1990, 50: 125-133 Symposium "Recovery from brain damage: behavioral and neurochemical approaches" 4-7 July, 1989, Warsaw, Poland ELECTROPHYSIOLOGICAL CORRELATES OF RECOVERY OF FUNCTION David S. OLTON and Matthew L. SHAPIRO Department of Psychology, Johns Hopkins University, 34th Charles St., Baltimore, Maryland 21218, USA and Department of Psychology, McGill University, Montreal, Quebec, H3A 1B1, Canada Key words: recovery of function, restoration, reorganization, hippocampus, grafts, place fields Abstract. The present experiment was designed to determine the extent to which two different mechanisms were responsible for the recovery of function produced by grafts of fetal basal forebrain tissue into the hippocampus of rats given fimbria-fornix lesions. The place fields of single units in the hippocampus were recorded as rats moved about a radial arm maze. Two measures of the discriminative stimuli identifying the place fields were determined: place stability, maze stability. In control (CON) rats, place stability was high and maze stability was low, indicating that the relevant discriminative stimuli were extramaze cues. In rats with fimbria-fornix (FF) lesions, place stability was low and maze stability was high, indicating that the relevant discriminative stimuli came from the maze itself. In rats given grafts of fetal basal forebrain tissue (GRAFT), place stability was increased and maze stability was reduced as compared to FF rats. These results indicate that the grafts produced recovery of function through restoration of normal hippocampal unit activity, rather than by reorganization of mew neural networks.
RESTORATION AND REORGANIZATION Many different mechanisms may underlie recovery of function following brain damage (3). Two of these are particularly important for the present experiment. Restoration refers to processes that help return the system to its original configuration, and produce recovery of function by enabling the system to use its normal mechanisms in a more effective manner. Reorganization refers to mechanisms that bypass the damaged ones, and produce recovery of function through alternative mechanisms, ones that are not normally used by the intact system. Distinguishing between reorganization and restoration cannot be accomplished easily without detailed analysis of the individual components of the system. Many different mechanisms are capable of transforming any given input to any given output, and an analysis from outside the system is usually not capable of determining which transformation is actually used. Rather, individual components of the system must be measured directly. This logic applies at each level of neuronal organization, from biophysical properties of membranes to systems of networks. In neural systems, electrophysiological recording from single units can be an effective means of determining the extent to which a neural network recovers through restoration or reorganization. If restoration underlies recovery of function, then the same parameters that affected single unit activity in the normal system should affect it in the recovered system. If, however, reorganization underlies recovery, then the variables that affect unit activity following recovery should be different than those that affected it in the normal system. In the experiments summarized here, the analysis of restoratian and reorganization was addressed by recording single unit activity from hippocampal cells in three groups of rats (16). Control (CON) rats were normal. Fimbria-fornix (FF) rats had lesions of the fimbria-fornix. Graft rats (GRAFT) has lesions of the fimbria-fornix and intrahippocampal grafts of fetal basal forebrain tissue. The activity of single units was recorded from cells in the hippocampus as rats moved about a spatial environment. An impairment of spatial memory was produced by FF lesions, recovery of function was produced by intrahippocampal grafts of fetal tissue frm the basal forebrain, and reoordings were made from single units in the hippocampus (13, 16). This combination of procedures is appropriate because FF lesions impair spatial memory and alter the parameters that affect single unit activity in the hippocampus (8), and intrahippocampal grafts of fetal tissue from the basal forebrain produce behavioral recovery of function. If this behavioral recovery is mediated
by restoration, then the variables that influence the activity of hippocarnpal single units in rats with grafts should be more similar to tho= variables in normal rats than in rats with FF lesions. If this behavioral recovery is mediated by reorganization, then the variables that influence the activity of hippocampal single units in rats with grafts should not be normalized, and may actually become more divergent from those affecting the activity of hippocampal units in normal rats. PLACE FIELDS One major variable influencimng the activity d many hippocampal complex spike units is the location of the rat in the environment (6, 12, 14, 15). The place field of a unit is ain area of the environment where the rate of single unit activity is markedly increased as compared to the rate in the rest of the environment. Two measures of single unit activity assessed the characteristics of the place field of complex spike units in the hippocampus. These measures were obtained by rotating a test maze in a room so that each part of the mlaze was rotated to the location of some other part of the maze prior to rotation. (1) Place stability measured the extent to which the place field of the unit remained in the same looation in the room after rotation as before rotation (and cmsequently was on a different part of the maze). High place stability indicates that the location of the place field was determined by extramaze stimuli, spatial stimuli that were located in the room outside of the maze. (2) Maze stability measured the extent to' which the place field was located in the same part of the maze after rotation as befwe rotation (and consequently was in a different location in the roam). High maze stability indictates that the locatim of the place field was influenced by intramaze stimuli, stimuli that are located m the apparatus itself. The characteristics of place fields diffe~ed in CON rats and FF rats. CON rats had high place stability and low maze stability, indicating that the relevant discriminative stimuli for the place field were extramaze. FF rats had high maze stability and low place stability, indicating that the discriminative stimuli for the place fields were intramaze. This distinction provided the basis to assess the mechanism by which grafts produce behavional recovery of function. Restoratiom would be indicated by n~rrnalizati~m of the characteristics of the place fields in the GRAFT group, an increase in place stability and a decrease in maze stability as compared to the FF group. Reorganization wmld be indicated by some other pattern, either no change from the FF group, or perhaps a more extreme deviation from normal than in the FF group. The procedures amd results for these experiments have been desczi- 3 - Acta Neurobiol. Exp. 4-5/90
bed in detail elsewhere (13, 16) and thus will be reviewed only briefly here. Female Sprague-Dawley rats were divided into three groups. The control group (CON) received no surgery. The FF group was given aspiration lesions of the fimbria-fornix which completely removed the FF bilaterally. The GRAFT group was given FF lesions and an intrahippocampal graft of tissue taken from the basal fo~ebrain of rats 15-17 days after gestation. Most of the rats were tested in a spatial discrimination (13) designed to test their ability to locate a hidden platform and move directly towards it from several different starting locations (10, 11). The GRAFT rats used for electrophysiological recording in the present experiment had the best recovery in this previous test of spatial memory (10, 11). For electrophysiological recording, a bundle of electrodes was mounted so that it could be driven ventrally through the hippocampus (4). Signals from the electrodes were amplified and discriminated using standard techniques. A radial arm maze with eight arms was located in a recording chamber with many cues surrounding the maze. The location of the rat on the maze was detected by a matrix camera, which located the lights placed on tap of the rat's head. Information about unit activity and the rat's location was processed and stored by a microcmputer. Each rat was trained to go down each arm of the maze to obtain a drop of chocolate (milk at the end of each arm. At the beginning of each trial, a drop of chocolate milk was placed in the cup at the end of each arm. After the rat had gone d m an m, obtained the chocolate milk, and returned to the center platform, that arm was slid horizontally away from the central platform so the rat no longer had access to it. In this way, the rat enbred each arm once and only cmce during each trial, insuring an equal number of visits to each arm. For each subsequent trial, chocolate milk was placed at the end of each arm, all the arms were pushed back to the center platform, and the procedure repeaked. MAZE STABILITY, PLACE STABILITY Each recording session had two types of trials. For traversal trials, the procedure was as just described with the anms of the maze in their usual location. For rotated trials, the maze was rotated 90' clockwise so that a new arm was in each spatial location. Four traversal trials and two rotated trials were given in each session. For the present discussion, the two analyses described earlier provide %he critical data. Both of them measured the correlation of the pattern
of place-field activity in the traversal and rotated trials. Place stability is the correlation of the pattern of activity with respect to locations in the room, irrespective of the parts of the maze occupying those locations. Maze stability is the correlation of the patterns of activity in those two trials with respect to the particular parts of the maze, irrespective of the location of those parts in the room. As described above, high place stability indicates that place fields are influenced primarily by extramaze stimuli (outside of the maze in the recording roam), while high maze stability indicates that the discriminative stimuli affecting the location of the place fields come primarily from intrm'aze stimuli (on the maze itself). These two measures are not entirely independent (nor are they completely redundant). Any given value fcrr one type of stability places an upper limit, but not a lower limit, for the magnitude d the other type of stability. The results af this analysis are presented in Fig. 1. Both axes quantify the respective correlations in terms of an r-to-z transformation to compensate for the skewed distribution of data. The vertical axis is place stability, the hmizontal axis is maze stability. For each of the three groups, the intersection of the two lines indicates the mean value along each dimension. The length of the line along the appropriate dimension indicates the standard error of the mean. Place- and Maze-Field Stability Differed Between Groups a I.L.. Fig. 1. A summary of the data from b hippocampal single units in three i +. GRAFT groups of rats: CON (control); FF I + (fimbria-fornix lesions); GRAFT (intrahipcampal graft of fetal basal + F F forebrain tissue after fimbria-fornix... : 4 lesion). For further explanation, see 0 0.1 1.2 1.3 0.4 1.5 0.6 text. Maze Stability CON rats had relatively high place stability and relatively low maze stability. FF rats had the opposite pattern of results, relatively high maze stability, and relatively low place stability. GRAFT rats showed
a marked decrease in maze stability, antd a slight increase in place stability. Both of these changes in the GRAFT rats were towards normalization of the characte~istics of place fields. CONCLUSIONS Restoration, rather than reorganization, is the most likely mechanism responsible for the behavioral recovery of function produced by grafts of fetal tissue from the basal forebrain in the hippocampus of adult rats with lesions of the fimbria-fornix (FF). Reoordings of single unit activity of hippotcairnpal complex spike cells in rats with grafts showed that the variables influencing the rate of activity of these units were normalized, more similar to those in control (CON) rats than in FF rats. The behavioral impairment produced by FF lesions has been described in detail elsewhere (13) and is consistent with interpretations emphasizing spatial memory functions of the hippocampus (6, 12, 14, 15). The behavimal improvement produced by the grafts is described in detail elsewhere (13). The recovery was not complete in all rats, and was marginal in some, indicating substantial individual differences in the extent to which the grafting procedure produced recovery of function. Interpreting this variability is currently still difficult. It may reflect technical limitations of the current procedures, which can be resolved with further improvement of these methods. Alternatively, it may reflect the fact that these particular grafts did not completely rectify all of the damage induced by the FF lesion. For example, cells that send afferents to or receive effe~ents from the hippocampus remained disconnected from the hippocampus after these intrahippcampal grafts. Within the hippocampus, the graft did not provide complete innervation throughout the hippocampus by all of the neural components normally present in the FF. Thus, the partial recovery produced by the graft may reflect the maximum recovery of function that can be attained with this particular intervention. The present experiment was not designed to address these issues, which will have to be pursued in other work. For the analysis pursued here, complete behavioral recovery is not necessary, and the partial improvements p~oduced by the grafts are sufficient to examine the associated changes in electrophysiological activity of the hippocampus. The fact that partial recovery of behavior and of electrophysiology took place in spite of the continued absence of both afferent and efferent fibers in the FF indicates that these fibers are not critical for at least the level of recovery observed here, implying that reciprocal connections of the hippocampus with the rest of the brain through the entorhinal
cortex must have been involved in the recovery (2, 7). Interpreting the role of entorhinal connections in the remaining electrophysiological and behavioral impairments is more difficult (9). The failure to obtain a complete recovery might indicate that the FF fibers are necessary to normalize completely electrophysiology and behavior. Alternatively, the absence of complete recovery might reflect technical limitations in the ability of current grafting techniques to produce optimal innervation of the host tissue. A comparison of the effects of different kinds of grafts, especially ones that differ in their connectivity, can help answer these questions (1). Further quantitative analyses of the relative magnitude of changes in maze stability and place stability are also desirable to determine the extent to which these two changes are independent of each other. Quantitatively, grafts produced a greater decrease in maze stability than an increase in place stability; they normalized maze stability more than place stability. This difference in the magnitude of change might indicate that measures of maze stability are more sensitive than measures of place stability to the beneficial effects of grafts. If such is the case, then changes in maze stability and place stability should be highly correlated with each other with a positively accelerated function. Alternatively, the difference in magnitude might indicate some independence of mechanisms such that reduction of abnormal correlates of unit activity (maze stability) may take place through a mechanism different than improvements in normal correlates (place stability). An independence of these two mechanisms should be indicated in a relatively low correlation of graft-induced changes in maze stability and place stability. Ultimately, of course, a thorough assessment of the many different types of electrophysiological characteristics of the hippocampus following grafts will be required to indicate the extent to which different changes occur independently. Because the present experiment used a between-subjects design, several different types of neural changes might underlie the differences in place fields in the various groups. If the same population of neurons was sampled equivalently in all three experimental groups, then the differences amomg the groups indicate that the lesion and the grafts altered, in complementary ways, the activity of the same set of neurons. However, many different variables might have led to selective recording from different populations of neurons in one or more of these experimental groups. If such is the case, then the manipulation in that group changed some variable (overall excitability for example) of a subset of neurons which made recording from that subset more likely. Distinguishing among these and other alternative explanations of the differences
among the experimental groups is necessary to provide a detailed model of how lesions and grafts affect neurons in the hippocampus. For obvious reasons, this within-subjects design is inacredibly difficult because it requires recording from the same neuron for an extended period of time in spite of lesions and grafts, both of which produce physical distortion of the hain. Cornsequently, this detailed level of analysis may be very difficult to obtain. Nonetheless, the absence of this information does not compromise the interpretation at a less detailed level of analysis. FF lesions changed some neurons in the hippocampus (15), altering the variables that influence the place fields, and grafts helped restore some neurons with miore normal place fields. Thus, at the level of single unit activity, restoration, rather than reorganization, was the change associated with recovery of function. Whether restoration or reorganization is the associated mechanism for an individual neuron remains to be determined with the appropriate within-subjects analysis. D. S. Olton thanks Professor K. Zielinski and B. Oderfeld-Nowak for their kind hospitality and the invitation to attend this conference, and acknowledges constructive comments made by participants in that conference. This research was supported in part by the National Science Foundation (NSF 8412613) and by the National Institute of Aging (Alzheimer's Disease Research Center, 5P50AG05146 to D. S. Olton, and NIA AGO6088 to F. H. Gage). REFERENCES 1. BUZSAKI, G., CZOPF, J., KONDAKOR, I., BJORKLUND, A. and GAGE, F. H. 1987. Cellular activity of intracerebrally transplanted fetal hippocampus during behavior. Neuroscience 22 (3): 871-883. 2. DUNNETT, S. B., LOW, W. C., IVERSON, S. I)., STENEVI, U. and BJORK- LUND, A. 1982. Septa1 transplants restore maze learning in rats with fornix-fimbria lesions. Brain Res. 251: 335-348. 3. FINGER, S. and STEIN, D. G. 1982. Brain damage and recovery; research and clinical perspectives. Academic Press, New York. 4. KUBIE, J. 1984. A driveable bundle of microwires for collecting single unit data from freely moving rats. Physiol. Behav. 32: 115-118. 5. KUBIE, J. and RANCK, J. B. Jr. 1983. Sensory-behavioral correlates in individual hippocampus neurons in three situations: space and context. In W. Seifert (ed.), The neurobiology of the hippocampus. Academic Press, London. 6. LEONARD, B. J. and MCNAUGHTON, B. L. 1989. Spatial representation in the rat: conceptual, behavioral, and neurophysiological perspectives. In R. Kesner md D. S. Olton (ed.), The neurobiology of comparative cognition. Lawrence Earlbarn Assoc., Hillsdale, p. 363-422. 7. LOW, W. C., LEWIS, P. R., BUNCH, S. T., DUNNETT, S. B., THOMAS, S. R., IVERSON, S. D., BJORKLUND, A. and STENEVI, U. 1982. Functional recovery following neural transplantation of embryonic septa1 nuclei in adult rats with septohippocampal lesions. Nature 300: 260-262.
8. MILLER, R. V. M. and BEST, P. J. 1980. Spatial correlates of hippocampal unit activity are altered by lesions of the fornix and entorhinal cortex. Brain Res. 194: 311-323. 9. MIZUMORI, S. J. Y., BARNES, C. A., MCNAUGHTON, B, L. and FOX, K. B. 1989. Preserved spatial coding in hippocampal pyramidal cells during reversible suppression of CA3, memory disruption. J. Neurosci. 11: 3915-3928. 10. MORRIS, R. G. M. 1983. An attempt to dissociate "spatial mapping" and "working memory" theories of hippocampal function. In W. Seifert (ed.), The neurobiology of the hippocampus. Academic Press, London. 11. MORRIS, R. G. M. 1984. Developments of a water-maze procedure for studying spatial learning in the rat. J. Neurosci. Meth. 11: 47-60. 12. MULLER, R. U., KUBIE, J. L. and RANCK, J. B. Jr. 1987. Spatial firing patterns of hippocampal complex-spike cells in a fixed environment. J. Neurosci. 7 (7): 173-195. 13. NILSSON, O., SHAPIRO, M. L., GAGE, F. H., OLTON, D. S. and BJ~RK- LUND, A. 1987. Spatial learning and memory following fimbria-fornix transection and grafting of fetal septal neurons to the hippocampus. Exp. Brain Res. 67: 195-215. 14. O'KEEFE, W. and NADEL, L. 1978. The hippocampus as a cognitive map. Clarendon Press, Oxford. 15. OLTON, D. S. 1989. Mnemonic functions of the hippocampus: single unit analyses in rats. In V. Chan-Palay and C. Kohler (ed.), The hippocampus-new vistas. A. R. Liss, New York, p. 411-424. 16. SHAPIRO, M. L., SIMON, D. K., OLTON, D. S., GAGE, F. H., NILSSON, 0. and BJORKLUND, A. 1989. Intrahippocarnpal graft of fetal basal forebrain tissue influence the place-correlates of complex-spike units in the hippacampus of behaving rats with fimbria-fornix lesions. IBRO Neurosci. 32: 1-18.