JOURNAL OF NEUROTRAUMA Volume 8, Number 4, 1991
Mary Ann Liebert, Inc.,
Publishers
Evaluation of Memory Dysfunction Following Experimental Brain Injury Using the Morris Water Maze DOUGLAS H. SMITH, KOICHI OKIYAMA, MARK J. THOMAS, BRIAN CLAUSSEN, and TRACEY K. McINTOSH
ABSTRACT
Memory dysfunction, a common clinical feature of traumatic brain injury (TBI), is thought to be related to secondary damage of key anatomic structures in the brain, including the hippocampus. In the present study, we have characterized and evaluated a novel experimental paradigm using the Morris water maze (MWM) technique, to measure post-TBI memory retention after lateral (parasagittal) fluid percussion (FP) brain injury in rats. Male Sprague-Dawley rats (n 37) received a total of 20 training trials over 2 days in the MWM. Two and a half hours after the last training trial, the animals received FP brain injury of moderate severity (2.3 atmospheres, n 12), high severity (2.6 atm, n 13), or no injury FP brain hours after injury, we observed a highly sufficient memory (n 12). Forty-two from both animals in injury groups compared to the uninjured group dysfunction of The degree memory dysfunction was found to be directly related to the severity (p < 0.001). of injury, with the high severity group scoring significantly worse than the moderately injured group (p 0.15). In addition, hippocampal cell loss was observed after brain injury, but only unilaterally. These data suggest that lateral FP brain injury causes memory dysfunction possibly related to concurrent hippocampal cell loss and that posttraumatic memory deficits may be sensitively quantitated using the memory testing paradigm described. =
=
=
=
=
INTRODUCTION memory dysfunction is one of the most debilitating and distressing features following traumatic brain injury (TBI) (Parkin, 1984; Levin, 1985). Despite this, few experimental models have been characterized that measure specific components of post-TBI memory dysfunction. A study by Lyeth et al. (1990) demonstrated a learning/memory impairment following midline fluid percussion (FP) brain injury (using the eight-arm radial maze). However, no experimental model has been developed that examines the effects of brain injury on memory as a distinct entity. Many clinical and experimental studies have shown that bilateral damage to key structures in the brain, including the hippocampus and other limbic structures,
Clinically,
CNS Injury Laboratory, Surgical Research Center, Department of Surgery, University of Connecticut Health Center, Farmington, Connecticut.
259
SMITH ET AL. may
in the development of memory disorders (Scoville and Milner, 1957; Zola-Morgan 1986). These studies have shown the hippocampus, in particular, to be an integral modulator of memory
play a major role
et al.,
function. It has further been proposed that the hippocampus is involved selectively with spatial learning and memory (Olton et al., 1978; Morris et al., 1982). Histologie studies have demonstrated that specific regions of the hippocampus (CA-1 and CA-3) appear to be exquisitely vulnerable to a variety of central nervous system (CNS) insults, including ischemia (Briefly, 1976; Pulsinelli et al., 1982), hypoglycemia (Aueret al., 1984), and trauma (Cortez et al., 1989), showing disproportionately high neuronal cell loss compared to that of adjacent structures. The systematic evaluation of memory function after TBI may offer a unique approach for the assessment of potentially related hippocampal function in an in vivo model. In addition, the development of a reproducible memory test would facilitate a more clinically relevant evaluation of post-TBI pharmacologie therapies, since encouraging histologie studies have shown that post-CNS injury treatment with various pharmacologie agents can attenuate hippocampal cell loss (Simon et al., 1984; Swan and Meldrum, 1990). In the present study, we characterized and evaluated a novel experimental paradigm using the Morris water maze (MWM) (Morris, 1984) to assess post-TBI memory deficits induced by lateral FP brain injury in rats.
MATERIALS AND METHODS Morris Water Maze Our model of the MWM is a circular pool 1 meter in diameter and 50 cm deep constructed of aluminum sides and a Plexiglas and wood bottom. The interior of the maze is painted white. The pool is filled with 18°C tap water to 25 cm in depth. A grid design of various derived zones (Fig. 1) is constructed with a computerized video system (Omnitech Videoscan) and is superimposed over the maze and viewed on a monitor. A platform 24 cm tall (1 cm below the surface of the water) and 11.5 x 11.5 cm is placed into the tank so that zone A
FIG. 1. Computerized grid design superimposed over the MWM during memory testing. Discrete zones are labeled with block letters, zone A representing the platform site. Animal swimming patterns (activity patterns) and time (seconds) spent in each zone are recorded during each memory test.
260
EFFECTS OF BRAIN INJURY ON MEMORY FUNCTION
(viewed on monitor) is completely within its borders. The
maze is then filled with small granular pieces of the surface of the is rendered water that styrofoam totally opaque. An alternative for making the maze mix milk with or is the but we found this method to be less optimal. The maze to water, powdered opaque paint is situated in a room with a variety of external visual cues easily seen from inside the tank, which remain constant throughout training and testing.
so
Training Male Sprague-Dawley rats, 350-400 g (n 37), were examined for corneal opacities or other ocular defects. Only those animals with no apparent visual defects or potential health disorders (e.g., exúdate around eyes, poor grooming) were selected. Animals were placed into the maze by holding them behind the front legs and gliding them into the water so that their noses are brought up to side of the tank at a randomly chosen site. The essential feature of the MWM is that the animals can escape from the water onto the submerged platform. The platform is rendered invisible, thus offering no local cues to guide escape behavior. The experimenter handling the animals stands back on a mark 1.5 meters from the maze after placing the animals in the water, while observing their swimming behavior on the monitor. For each training trial, the animals were placed randomly at four sites, 90 degrees apart, along the tank's periphery. On the first day of training (afternoon), each animal was given 2 minutes to find the platform (approximately one third of the animals did not find the platform on their first trial and were placed on it once the 2 minutes had elapsed). The animals were allowed to remain on the platform for 30 seconds on their first trial and 15 seconds on subsequent trials to spatially orient themselves to the external visual cues. Each animal received a total of 10 trials on day 1 of training, and the latencies (time taken to find the platform) were recorded for each trial. Once training was completed, animals were returned to the animal room overnight. The following morning, all animals were put through the identical training regimen of 10 trials, for a total of 20 trials over 2 days. The average of an individual animal's latency scores to find the platform after two sets of training was typically 4-6 seconds. =
Fluid Percussion Brain
injury
At 2.5 hours after completion of training in the MWM, all animals were anesthetized with sodium pentobarbital (60 mg/kg) 5 minutes after receiving 0.15 ml of atropine (0.4 mg/ml). For both FP brain injury and sham (uninjured) groups, animals were placed in a stereotaxic head holder, and ophthalmic ointment was applied to their eyes to protect vision during surgery. The top of the skull was exposed, and a craniotomy was performed over the left parietal cortex centered between bregma and lambda. A Luer-Lok hub was fixed to the craniotomy site with dental cement. At 1.5 hours after administration of anesthesia, the animals were subjected to FP brain injury of either moderate (2.3 atm, n 12) or high severity (2.6 atm, n 13). Sham-operated (uninjured) animals (n 12) served as controls. The FP device consisted of a Plexiglas, cylindrical, saline-filled reservoir bounded at one end by a Plexiglas plunger mounted on O-rings. The opposite end of the cylinder was capped with a male Luer stub, which connected to the Luer-Lok cemented on the craniotomy. A pendulum was allowed to drop, striking the plunger and causing a rapid and high pressure injection of saline into the closed cranial cavity, compressing the brain. The injury level (atm) was measured by a pressure transducer recorded on an oscilloscope. This model produced a reproducible degree of brain injury, including focal injury in the left parietotemporal cortex. The lateral FP brain injury method has been =
=
=
described in detail (Mclntosh et al., 1989b).
Postinjury Memory and Latency Tests All postinjury memory tests were conducted at 42 hours after FP brain injury (48 hours after the last training session), since animals tested at later timepoints demonstrated some extinction of this newly learned task. With the platform now removed, each animal was given two 1-minute tests in the maze, and the following behavioral indices were recorded: (1) activity or swimming pattern, (2) time spent in each zone, and (3) latency (first crossing through zone A), using the Omnitech Videoscan zonal behavior program. 261
SMITH ET AL. Each zone of the computerized grid design was ranked in a weighted fashion according to its proximity to the platform site. These assigned numbers were multiplied by the number of seconds spent in the corresponding zone and totaled. The resulting score for each animal's first 1-minute trial was designated ml, and the score from the second 1-minute trial was designated m.2. The ranked numbers shown in Figure 1 were derived from a series of pilot studies designed to assess the behavioral characteristics of trained animals searching for the platform. In these studies, animals who received a more severe brain injury, when tested for memory, typically swam along the periphery of the tank and only occasionally exhibited seeking behavior by swimming through the center of the tank. However, uninjured sham animals or animals receiving brain injury of mild severity typically exhibited more seeking behavior by swimming in straight or curved trajectories through or near the platform site (zone A), stopping to survey their surroundings once past it, and swimming back through again. Because of its combined position in the periphery and proximity to the platform site, zone D was scored as a function of the time spent in zone D divided by the time spent in zone G (peripheral zone,
Fig. 1).
Three
general types of behavior were demonstrated by individual animals during their two memory tests:
(1) animals who scored very highly during ml and lower during m2 (possibly due to extinction of the behavior), (2) animals who scored poorly during 1 ml, but greatly improved their seeking behaviors and corresponding scores during m2 (possibly due to acclimation problems during their ml), and (3) animals who did similarly in both memory tests. Therefore, a scoring system was developed so that animals with ml > m2
assigned only their m 1 score, whereas animals with m2 > m 1 were assigned the average of their m 1 and m2 scores. The actual scoring for each memory trial is outlined in Figure 1. Statistical analyses for the measurement of memory dysfunction were performed using the Mann-Whitney U test. Latency scores were determined using the best latency times from ml or m2. Statistical analyses for the latency scores were performed using the Student's Mest. scores are
Histology Animals (n 12) used for histologie evaluation were injured at a moderate level (2.3 atm, n 6), high level (2.6 atm, n 6), or no injury (n 6). Forty-eight hours after brain injury, animals were reanesthetized (pentobarbital, 60 mg/kg) and perfused with 4% paraformaldehyde. The brains were removed and, after immersion in fixative, were cut into 50 p.m coronal sections on a Vibratome. Sections were stained with 1% toluidine blue (staining for Nissl granules and nuclei) according to the protocol of Disbrey and Rack (1970). Microscopic examination of the stained sections was performed for gross determination of neuronal cell loss (absence of Nissl staining) in the hippocampus. =
=
—
=
RESULTS A consistently large disparity in memory test activity patterns was observed between injured and uninjured animals. The range limits of the activity patterns are demonstrated by four examples shown in Figure 2. The calculated memory scores reflect and numerically define the differences in activity patterns of each animal. Statistical analyses comparing memory scores of each group revealed that animals receiving FP brain injury of moderate severity showed a highly significant memory deficit (median score of 55) compared to uninjured animals (median score of 127, p < 0.001). Similarly, animals who received FP brain injury of high severity also showed a highly significant memory deficit (median score of 23) when compared to uninjured control animals (p < 0.001). A significant difference was observed between the memory scores of the moderate injury group and the high injury group (p = 0.015), suggesting that the degree of memory dysfunction accurately reflected the severity of brain injury (Fig. 3). Mean latency scores in the moderately injured group (5.1 seconds) and high injury group (37 seconds) were significantly higher compared to those of uninjured control animals (3.25 seconds, p < 0.05 and/? < 0.001, respectively). There also was a highly significant difference in latency scores between both injury groups (p < 0.001), demonstrating, as with the memory scores, an injury dose-response effect corresponding to the severity of injury (Fig. 4).
262
EFFECTS OF BRAIN INJURY ON MEMORY FUNCTION
FIG. 2. Activity patterns of four animals during a 1-minute memory test, demonstrating the range limits of uninjured animals (A,B) and animals receiving FP brain injury of high severity (2.6 atm) (C,D).
Three representative activity patterns reflecting memory scores near the median of each injury group are shown in Figure 5. These patterns visually demonstrate a graded effect, relative to the severity of FP brain injury, of an individual animal's seeking behavior of the platform site (zone A, Fig. 1) during a 1-minute memory test. Histologie evaluation performed 48 hours postinjury on animals who received moderate or high injury demonstrated a loss of neurons in the CA-2 and CA-3 regions of the left hippocampus (ipsilateral to injury site) compared to uninjured control animals. This pattern of cell loss in the CA-2 and CA-3 regions remained consistent throughout most of the rostral-caudal extent of the hippocampus, involving both ventral and dorsal tiers. With the methods used, there was no clear difference in the degree or location of cell loss between animals injured at moderate or high levels. The right hippocampus of injured animals showed no grossly discernable cell loss. These results were consistent in all the animals evaluated. A comparison of coronal sections of representative left hippocampus from an uninjured and an injured (moderate) animal is seen in Figure 6.
DISCUSSION The present study demonstrates that parasagittal FP brain injury causes profound spatial memory dysfunction, which may be sensitively measured using the MWM. Posttraumatic deficits in memory function were found to be related directly to the severity of brain injury. Histopathologic evaluation revealed marked hippocampal cell loss in the CA-2 and CA-3 regions, which may be related to this posttraumatic memory dysfunction. Our adaptation of the MWM is designed to examine memory alone, as opposed to the more usual learning/memory paradigms, thus simplifying possible interpretations of the recorded data. In addition, most learning/memory paradigms use only latency sources as an outcome parameter. We found in our model that
263
SMITH ET AL.
MEMORY SCORE
200
I UNINJURED CONTROL ANIMALS CXI INJURED ANIMALS 2.3 ATM EZ3 INJURED ANIMALS 2.6 ATM
j
175150CO
I-
125-
z
O CL.
100-
_l
Mary Ann Liebert, Inc.,
Publishers
Evaluation of Memory Dysfunction Following Experimental Brain Injury Using the Morris Water Maze DOUGLAS H. SMITH, KOICHI OKIYAMA, MARK J. THOMAS, BRIAN CLAUSSEN, and TRACEY K. McINTOSH
ABSTRACT
Memory dysfunction, a common clinical feature of traumatic brain injury (TBI), is thought to be related to secondary damage of key anatomic structures in the brain, including the hippocampus. In the present study, we have characterized and evaluated a novel experimental paradigm using the Morris water maze (MWM) technique, to measure post-TBI memory retention after lateral (parasagittal) fluid percussion (FP) brain injury in rats. Male Sprague-Dawley rats (n 37) received a total of 20 training trials over 2 days in the MWM. Two and a half hours after the last training trial, the animals received FP brain injury of moderate severity (2.3 atmospheres, n 12), high severity (2.6 atm, n 13), or no injury FP brain hours after injury, we observed a highly sufficient memory (n 12). Forty-two from both animals in injury groups compared to the uninjured group dysfunction of The degree memory dysfunction was found to be directly related to the severity (p < 0.001). of injury, with the high severity group scoring significantly worse than the moderately injured group (p 0.15). In addition, hippocampal cell loss was observed after brain injury, but only unilaterally. These data suggest that lateral FP brain injury causes memory dysfunction possibly related to concurrent hippocampal cell loss and that posttraumatic memory deficits may be sensitively quantitated using the memory testing paradigm described. =
=
=
=
=
INTRODUCTION memory dysfunction is one of the most debilitating and distressing features following traumatic brain injury (TBI) (Parkin, 1984; Levin, 1985). Despite this, few experimental models have been characterized that measure specific components of post-TBI memory dysfunction. A study by Lyeth et al. (1990) demonstrated a learning/memory impairment following midline fluid percussion (FP) brain injury (using the eight-arm radial maze). However, no experimental model has been developed that examines the effects of brain injury on memory as a distinct entity. Many clinical and experimental studies have shown that bilateral damage to key structures in the brain, including the hippocampus and other limbic structures,
Clinically,
CNS Injury Laboratory, Surgical Research Center, Department of Surgery, University of Connecticut Health Center, Farmington, Connecticut.
259
SMITH ET AL. may
in the development of memory disorders (Scoville and Milner, 1957; Zola-Morgan 1986). These studies have shown the hippocampus, in particular, to be an integral modulator of memory
play a major role
et al.,
function. It has further been proposed that the hippocampus is involved selectively with spatial learning and memory (Olton et al., 1978; Morris et al., 1982). Histologie studies have demonstrated that specific regions of the hippocampus (CA-1 and CA-3) appear to be exquisitely vulnerable to a variety of central nervous system (CNS) insults, including ischemia (Briefly, 1976; Pulsinelli et al., 1982), hypoglycemia (Aueret al., 1984), and trauma (Cortez et al., 1989), showing disproportionately high neuronal cell loss compared to that of adjacent structures. The systematic evaluation of memory function after TBI may offer a unique approach for the assessment of potentially related hippocampal function in an in vivo model. In addition, the development of a reproducible memory test would facilitate a more clinically relevant evaluation of post-TBI pharmacologie therapies, since encouraging histologie studies have shown that post-CNS injury treatment with various pharmacologie agents can attenuate hippocampal cell loss (Simon et al., 1984; Swan and Meldrum, 1990). In the present study, we characterized and evaluated a novel experimental paradigm using the Morris water maze (MWM) (Morris, 1984) to assess post-TBI memory deficits induced by lateral FP brain injury in rats.
MATERIALS AND METHODS Morris Water Maze Our model of the MWM is a circular pool 1 meter in diameter and 50 cm deep constructed of aluminum sides and a Plexiglas and wood bottom. The interior of the maze is painted white. The pool is filled with 18°C tap water to 25 cm in depth. A grid design of various derived zones (Fig. 1) is constructed with a computerized video system (Omnitech Videoscan) and is superimposed over the maze and viewed on a monitor. A platform 24 cm tall (1 cm below the surface of the water) and 11.5 x 11.5 cm is placed into the tank so that zone A
FIG. 1. Computerized grid design superimposed over the MWM during memory testing. Discrete zones are labeled with block letters, zone A representing the platform site. Animal swimming patterns (activity patterns) and time (seconds) spent in each zone are recorded during each memory test.
260
EFFECTS OF BRAIN INJURY ON MEMORY FUNCTION
(viewed on monitor) is completely within its borders. The
maze is then filled with small granular pieces of the surface of the is rendered water that styrofoam totally opaque. An alternative for making the maze mix milk with or is the but we found this method to be less optimal. The maze to water, powdered opaque paint is situated in a room with a variety of external visual cues easily seen from inside the tank, which remain constant throughout training and testing.
so
Training Male Sprague-Dawley rats, 350-400 g (n 37), were examined for corneal opacities or other ocular defects. Only those animals with no apparent visual defects or potential health disorders (e.g., exúdate around eyes, poor grooming) were selected. Animals were placed into the maze by holding them behind the front legs and gliding them into the water so that their noses are brought up to side of the tank at a randomly chosen site. The essential feature of the MWM is that the animals can escape from the water onto the submerged platform. The platform is rendered invisible, thus offering no local cues to guide escape behavior. The experimenter handling the animals stands back on a mark 1.5 meters from the maze after placing the animals in the water, while observing their swimming behavior on the monitor. For each training trial, the animals were placed randomly at four sites, 90 degrees apart, along the tank's periphery. On the first day of training (afternoon), each animal was given 2 minutes to find the platform (approximately one third of the animals did not find the platform on their first trial and were placed on it once the 2 minutes had elapsed). The animals were allowed to remain on the platform for 30 seconds on their first trial and 15 seconds on subsequent trials to spatially orient themselves to the external visual cues. Each animal received a total of 10 trials on day 1 of training, and the latencies (time taken to find the platform) were recorded for each trial. Once training was completed, animals were returned to the animal room overnight. The following morning, all animals were put through the identical training regimen of 10 trials, for a total of 20 trials over 2 days. The average of an individual animal's latency scores to find the platform after two sets of training was typically 4-6 seconds. =
Fluid Percussion Brain
injury
At 2.5 hours after completion of training in the MWM, all animals were anesthetized with sodium pentobarbital (60 mg/kg) 5 minutes after receiving 0.15 ml of atropine (0.4 mg/ml). For both FP brain injury and sham (uninjured) groups, animals were placed in a stereotaxic head holder, and ophthalmic ointment was applied to their eyes to protect vision during surgery. The top of the skull was exposed, and a craniotomy was performed over the left parietal cortex centered between bregma and lambda. A Luer-Lok hub was fixed to the craniotomy site with dental cement. At 1.5 hours after administration of anesthesia, the animals were subjected to FP brain injury of either moderate (2.3 atm, n 12) or high severity (2.6 atm, n 13). Sham-operated (uninjured) animals (n 12) served as controls. The FP device consisted of a Plexiglas, cylindrical, saline-filled reservoir bounded at one end by a Plexiglas plunger mounted on O-rings. The opposite end of the cylinder was capped with a male Luer stub, which connected to the Luer-Lok cemented on the craniotomy. A pendulum was allowed to drop, striking the plunger and causing a rapid and high pressure injection of saline into the closed cranial cavity, compressing the brain. The injury level (atm) was measured by a pressure transducer recorded on an oscilloscope. This model produced a reproducible degree of brain injury, including focal injury in the left parietotemporal cortex. The lateral FP brain injury method has been =
=
=
described in detail (Mclntosh et al., 1989b).
Postinjury Memory and Latency Tests All postinjury memory tests were conducted at 42 hours after FP brain injury (48 hours after the last training session), since animals tested at later timepoints demonstrated some extinction of this newly learned task. With the platform now removed, each animal was given two 1-minute tests in the maze, and the following behavioral indices were recorded: (1) activity or swimming pattern, (2) time spent in each zone, and (3) latency (first crossing through zone A), using the Omnitech Videoscan zonal behavior program. 261
SMITH ET AL. Each zone of the computerized grid design was ranked in a weighted fashion according to its proximity to the platform site. These assigned numbers were multiplied by the number of seconds spent in the corresponding zone and totaled. The resulting score for each animal's first 1-minute trial was designated ml, and the score from the second 1-minute trial was designated m.2. The ranked numbers shown in Figure 1 were derived from a series of pilot studies designed to assess the behavioral characteristics of trained animals searching for the platform. In these studies, animals who received a more severe brain injury, when tested for memory, typically swam along the periphery of the tank and only occasionally exhibited seeking behavior by swimming through the center of the tank. However, uninjured sham animals or animals receiving brain injury of mild severity typically exhibited more seeking behavior by swimming in straight or curved trajectories through or near the platform site (zone A), stopping to survey their surroundings once past it, and swimming back through again. Because of its combined position in the periphery and proximity to the platform site, zone D was scored as a function of the time spent in zone D divided by the time spent in zone G (peripheral zone,
Fig. 1).
Three
general types of behavior were demonstrated by individual animals during their two memory tests:
(1) animals who scored very highly during ml and lower during m2 (possibly due to extinction of the behavior), (2) animals who scored poorly during 1 ml, but greatly improved their seeking behaviors and corresponding scores during m2 (possibly due to acclimation problems during their ml), and (3) animals who did similarly in both memory tests. Therefore, a scoring system was developed so that animals with ml > m2
assigned only their m 1 score, whereas animals with m2 > m 1 were assigned the average of their m 1 and m2 scores. The actual scoring for each memory trial is outlined in Figure 1. Statistical analyses for the measurement of memory dysfunction were performed using the Mann-Whitney U test. Latency scores were determined using the best latency times from ml or m2. Statistical analyses for the latency scores were performed using the Student's Mest. scores are
Histology Animals (n 12) used for histologie evaluation were injured at a moderate level (2.3 atm, n 6), high level (2.6 atm, n 6), or no injury (n 6). Forty-eight hours after brain injury, animals were reanesthetized (pentobarbital, 60 mg/kg) and perfused with 4% paraformaldehyde. The brains were removed and, after immersion in fixative, were cut into 50 p.m coronal sections on a Vibratome. Sections were stained with 1% toluidine blue (staining for Nissl granules and nuclei) according to the protocol of Disbrey and Rack (1970). Microscopic examination of the stained sections was performed for gross determination of neuronal cell loss (absence of Nissl staining) in the hippocampus. =
=
—
=
RESULTS A consistently large disparity in memory test activity patterns was observed between injured and uninjured animals. The range limits of the activity patterns are demonstrated by four examples shown in Figure 2. The calculated memory scores reflect and numerically define the differences in activity patterns of each animal. Statistical analyses comparing memory scores of each group revealed that animals receiving FP brain injury of moderate severity showed a highly significant memory deficit (median score of 55) compared to uninjured animals (median score of 127, p < 0.001). Similarly, animals who received FP brain injury of high severity also showed a highly significant memory deficit (median score of 23) when compared to uninjured control animals (p < 0.001). A significant difference was observed between the memory scores of the moderate injury group and the high injury group (p = 0.015), suggesting that the degree of memory dysfunction accurately reflected the severity of brain injury (Fig. 3). Mean latency scores in the moderately injured group (5.1 seconds) and high injury group (37 seconds) were significantly higher compared to those of uninjured control animals (3.25 seconds, p < 0.05 and/? < 0.001, respectively). There also was a highly significant difference in latency scores between both injury groups (p < 0.001), demonstrating, as with the memory scores, an injury dose-response effect corresponding to the severity of injury (Fig. 4).
262
EFFECTS OF BRAIN INJURY ON MEMORY FUNCTION
FIG. 2. Activity patterns of four animals during a 1-minute memory test, demonstrating the range limits of uninjured animals (A,B) and animals receiving FP brain injury of high severity (2.6 atm) (C,D).
Three representative activity patterns reflecting memory scores near the median of each injury group are shown in Figure 5. These patterns visually demonstrate a graded effect, relative to the severity of FP brain injury, of an individual animal's seeking behavior of the platform site (zone A, Fig. 1) during a 1-minute memory test. Histologie evaluation performed 48 hours postinjury on animals who received moderate or high injury demonstrated a loss of neurons in the CA-2 and CA-3 regions of the left hippocampus (ipsilateral to injury site) compared to uninjured control animals. This pattern of cell loss in the CA-2 and CA-3 regions remained consistent throughout most of the rostral-caudal extent of the hippocampus, involving both ventral and dorsal tiers. With the methods used, there was no clear difference in the degree or location of cell loss between animals injured at moderate or high levels. The right hippocampus of injured animals showed no grossly discernable cell loss. These results were consistent in all the animals evaluated. A comparison of coronal sections of representative left hippocampus from an uninjured and an injured (moderate) animal is seen in Figure 6.
DISCUSSION The present study demonstrates that parasagittal FP brain injury causes profound spatial memory dysfunction, which may be sensitively measured using the MWM. Posttraumatic deficits in memory function were found to be related directly to the severity of brain injury. Histopathologic evaluation revealed marked hippocampal cell loss in the CA-2 and CA-3 regions, which may be related to this posttraumatic memory dysfunction. Our adaptation of the MWM is designed to examine memory alone, as opposed to the more usual learning/memory paradigms, thus simplifying possible interpretations of the recorded data. In addition, most learning/memory paradigms use only latency sources as an outcome parameter. We found in our model that
263
SMITH ET AL.
MEMORY SCORE
200
I UNINJURED CONTROL ANIMALS CXI INJURED ANIMALS 2.3 ATM EZ3 INJURED ANIMALS 2.6 ATM
j
175150CO
I-
125-
z
O CL.
100-
_l
