- Email: [email protected]
Effect of age on the radial arm water maze—a test of spatial learning and memory
Neurobiology of Aging 25 (2004) 223–229
Effect of age on the radial arm water maze—a test of spatial learning and memory Barbara Shukitt-Hale∗ , John J. McEwen, Aleksandra Szprengiel, James A. Joseph USDA-ARS, Human Nutrition Research Center on Aging at Tufts University, 711 Washington Street, Boston, MA 02111, USA Received 9 October 2001; received in revised form 3 March 2002; accepted 21 February 2003
Abstract Aged rats show decrements in performance on cognitive tasks that require the use of spatial learning and memory. We used the 8-arm radial water maze (RAWM) to measure spatial learning as a function of age in young (6 months) and old (21 months) male F344 rats. Rats were placed in the RAWM in different start arms with the same goal arm for 3 days (five trials/day); the goal arm was changed on day 4. Old rats demonstrated spatial impairment as evidenced by increased latencies to find the hidden platform on day 4. Old rats made significantly more errors, both reference and working memory errors, than young rats on all days. It is likely that the old rats utilized non-spatial strategies to solve the task, and therefore were impaired in learning a new platform location. The RAWM is a reliable, sensitive, and powerful additional test to assess age-related spatial learning and memory deficits, combining the advantages of the Morris water maze and the radial arm maze while minimizing the disadvantages. © 2003 Elsevier Science Inc. All rights reserved. Keywords: Cognition; Reference memory; Working memory; Short-term memory; Long-term memory; Fischer 344 rat
1. Introduction Aged rats show decrements in performance on cognitive tasks that require the use of spatial learning and memory, that is, the ability to acquire a cognitive representation of location in space and the ability to effectively navigate the environment (for reviews see [2,6,13,16]). Memory alterations appear to occur primarily in secondary memory systems and are reflected in the storage of newly acquired information [3,17]. It is thought that the hippocampus mediates allocentric spatial navigation (i.e., place learning), and that the prefrontal cortex is critical to acquiring the rules that govern performance in particular tasks (i.e., procedural knowledge), while the dorsomedial striatum mediates egocentric spatial orientation (i.e., response and cue learning) [9,21,26,36]. Two well known tests used to assess spatial learning and memory in rats are the Morris water maze (MWM) [23,24] and the radial arm maze (RAM) [27,28]. Both of these tests have distinct advantages and disadvantages in assessing spatial learning and memory in rats. Therefore, in the present study, we used a combination of these two tests, the 8-arm radial water maze (RAWM), to measure spatial learning as a function of age in young ∗
Corresponding author. Tel.: +1-617-556-3118; fax: +1-617-556-3222. E-mail address: [email protected] (B. Shukitt-Hale).
and old rats. This test maintains the advantages of both the MWM and the RAM, while eliminating the disadvantages [8,10,14,15,30]. Several studies have used various versions of the RAWM to measure learning in different strains of mice [15], to assess sex differences in learning [4,5], to investigate the effects of stress on spatial working memory [10], and to assess learning and memory deficits that develop in a transgenic mouse model of Alzheimer’s disease [1,22]. The RAWM combines the variable spatial complexity of the RAM and the efficient learning of the MWM [10]. One distinct advantage of the RAWM is the ability to assess both reference and working memory errors simultaneously [15]. Reference memory is believed to reflect learning the trial-independent procedural aspects of the task (spatial cue locations), while working memory is trial-dependent and describes the ability of the subject to hold this trial-dependent information (places previously visited) in memory [11,20]. Furthermore, in the MWM the rat can use a strategy of swimming an appropriate fixed distance from the wall to find the hidden platform, while in the RAWM the rat is forced to swim in either the central open area or one of the arms, and therefore cannot solve the task by adopting this proximal strategy. The rat can also find the platform in the MWM by bumping into it, while this strategy is not possible in the RAWM without an error being assessed. Swimming around the sides of the pool is also not permitted in the RAWM. In
0197-4580/$ – see front matter © 2003 Elsevier Science Inc. All rights reserved. doi:10.1016/S0197-4580(03)00041-1
224
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
the RAM, food or water deprivation is required, while the RAWM is non-appetitively motivated and therefore testing in it is not confounded by differences in food motivation. Dietary interventions can also be tested in the RAWM since food deprivation is not necessary, and multiple trials can be run per day in the RAWM since satiation is not a factor. The effect of odor cues present in the RAM is eliminated in the RAWM. Also, since the rat is forced to swim in the RAWM, the rat cannot choose whether or not to move, as it can in the RAM, so failure to respond is not a confound. Thus, because of its advantages, we used the RAWM in the present study to measure spatial learning and memory as a function of age.
the platform. Once the rat reached the platform, it remained there for 15 s. At the end of each trial, the rat was towel-dried, returned to its home cage, and 30 min elapsed (during which the remaining rats were tested) before being returned to the maze for its next trial, which used the same platform location and a different start position as the previous trial. Performance on each trial was videotaped and analyzed with image tracking software (HVS Image, Hampton, UK), which provided dependent measures such as latency (s) and speed (cm/s). Also, the order of entry into the maze arms was recorded so that the number of errors could be analyzed; errors were broken into reference (long-term) memory (defined as entering an arm that does not contain the platform) and working (short-term) memory errors (reentries into an arm).
2. Materials and methods 2.3. Statistical analyses 2.1. Animals The subjects consisted of two age groups of male F344 rats (NIA colony, Harlan Sprague Dawley, Indianapolis, IN): young, 6 months (n = 10) and old, 21 months (n = 10), weighing between 300 and 440 g. The rats were individually housed in stainless steel mesh suspended cages, provided food and water ad libitum, and maintained on a 12 h light/dark cycle. All animals were observed daily for clinical signs of disease. The old animals weighed significantly more than the young animals (young: 333.30 ± 6.41 g (mean ± S.E.M.); old: 419.50 ± 6.01 g; F(1, 18) = 96.12, P < 0.001). 2.2. Apparatus The radial arm water maze consisted of a circular galvanized pool (152 cm diameter × 59 cm high) filled to a depth of 32 cm with water (23 ◦ C). Within the pool were stainless steel walls (40 cm high × 43 cm long), arranged to make eight swim paths or arms radiating from an open center area. The pool, walls, and circular escape platform (10 cm diameter × 30 cm high) were all colored black, enabling the platform to be hidden from sight. The maze was placed in a room with the lights dimmed, and there were numerous extramaze cues on the walls and scattered throughout the room that the rat could use to navigate the maze. On days 1–3, the platform was submerged 2 cm below the surface of the water and 4 cm from the end of arm 1, switching to arm 4 on day 4. RAWM testing was performed for 4 days, five trials/day with a 30 min intertrial interval. At the beginning of each trial, the rat was gently immersed in the water, facing the wall, at one of seven start positions. Start locations were randomized, with the qualifications that not the same starting location be used twice in a row and that the rat never be started in the arm with the platform. Each rat was allowed 120 s to escape onto the platform; if the rat failed to escape within this time, it was guided to
For each behavioral measure, repeated measures analysis of variance (ANOVA) models with one between-subjects factor (age) and one within-subjects factor (trials) were performed using Systat (SPSS, Inc., Chicago, IL) to test for statistical significance at the P ≤ 0.05 level. Separate one-way ANOVAs or t-tests were run to determine which trials were different between groups when a significant overall group difference or group × trial interaction was found. 3. Results Old rats demonstrated spatial learning and memory impairment in the RAWM, particularly on error measures and on the reversal day when the platform was moved. On days 1–3, there was significant improvement over time in water maze acquisition (learning) performance (i.e., latency to find the platform), F(4, 72) = 6.69, P < 0.001; improvements in performance were shown by both the old and young groups (Fig. 1). There was no difference in overall acquisition between groups on days 1–3, as latency measures were similar for the two age groups (P > 0.05). It was on day 4, when the location of the platform was changed from arm 1 to arm 4, that the old rats exhibited the most dramatic learning deficits as evidenced by increased latencies to find the hidden platform (subsequent to trial 1) (main effect of age: F(1, 18) = 3.94, P = 0.06; age × trial interaction: F(4, 72) = 3.29, P < 0.05; Fig. 1). When the platform location was moved on day 4, old rats initially found the new platform location more quickly (although not significantly) than the young on trial 1. The young rats, because they were using spatial strategies, took longer to find the new location since they initially searched longer in the former platform position, shown by more crossings of the old platform location (6.1 ± 1.01 crossings of the old platform location for young versus 3.00±0.84 for old rats; F(1, 18) = 5.58, P < 0.05). However, the young rats learned the new task faster than the old group, as the performance of the
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
225
Fig. 1. Escape latencies (s, mean ± S.E.M.) to find the hidden platform in the RAWM for days 1–3 (trials 1–5) and for reversal testing on day 4 (trials 1–5) for young (6 months) and old (21 months) F344 rats.
young improved by the second trial onwards, but the performance of the old group did not. By trials 4 and 5, the latency to find the platform for the young group was significantly less (P < 0.01) than the old group (Fig. 1). Old rats made significantly more errors than young rats on all days (Fig. 2). For days 1–3, ANOVA showed a significant age × trial interaction for total errors, F(14, 252) = 2.27, P < 0.01; although the main effect of age failed to reach statistical significance, F(1, 18) = 3.08, P = 0.09. More specifically, on day 1, young rats initially made more errors on trial 1 but quickly learned the task, F(4, 72) = 3.39, P < 0.05. On day 2, F(1, 18) = 6.13, P < 0.05; and day 3, F(1, 18) = 7.10, P < 0.05; there was a main effect of age, with the old rats making significantly more errors than
the young rats. On day 4, when the platform was moved, again the young rats initially made more errors (although not significantly more) before learning the new platform location, F(4, 72) = 2.58, P < 0.05; the old rats made significantly more errors on trials 2, 4, and 5 (P < 0.05) compared to the young group (Fig. 2). When errors were broken down into reference memory (Fig. 3A) and working memory errors (Fig. 3B), it was found that old rats made significantly more reference memory, F(14, 252) = 2.12, P < 0.05; and working memory, F(14, 252) = 2.18, P < 0.01; errors than young rats on days 1–3, as shown by significant age × trial interactions (again, however, the main effect of age failed to reach significance for reference memory, F(1, 18) = 2.89, P = 0.11;
Fig. 2. Total number of errors made (mean ± S.E.M.) while finding the hidden platform in the RAWM for days 1–3 (trials 1–5) and for reversal testing on day 4 (trials 1–5) for young (6 months) and old (21 months) F344 rats.
226
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
Fig. 3. Total number of reference memory (A) and working memory (B) errors made (mean ± S.E.M.) while finding the hidden platform in the RAWM for days 1–3 (trials 1–5) and for reversal testing on day 4 (trials 1–5) for young (6 months) and old (21 months) F344 rats. Reference memory errors are defined as entering an arm that does not contain the platform, while working memory errors are reentries into an arm.
Fig. 4. Speed (cm/s, mean ± S.E.M.) to find the hidden platform in the RAWM for days 1–3 (trials 1–5) and for reversal testing on day 4 (trials 1–5) for young (6 months) and old (21 months) F344 rats.
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
and working memory, F(1, 18) = 1.97, P = 0.18, errors). Old rats made more reference memory errors compared to young rats on day 2, F(1, 18) = 6.71, P < 0.05; day 3, F(1, 18) = 10.06, P < 0.01; and day 4, F(1, 18) = 9.35, P < 0.01; and more working memory errors on day 2, F(1, 18) = 4.07, P ≤ 0.05; and day 4, F(1, 18) = 6.68, P < 0.05; but not day 3. On day 1, young rats initially made more working memory errors but steadily improved over the trials on this day and did not reenter the arms as frequently as old rats in later trials, F(4, 72) = 3.91, P < 0.01. On days 2–4, young rats swam faster than old animals (day 2, F(1, 18) = 6.32, P < 0.05; day 3, F(1, 18) = 10.22, P < 0.01; day 4, F(1, 18) = 13.29, P < 0.01; Fig. 4). However, there was no difference in speed between the groups on day 1, ruling out gross motor impairment. Instead, the young animals swam faster as the trials progressed, while the old rats maintained one speed or did not increase their speed to the extent of the young, particularly on days 2 and 4 (day 2, F(4, 72) = 3.56, P < 0.05; day 4, F(4, 72) = 4.99, P < 0.01; Fig. 4).
4. Discussion In the RAWM, as in the MWM [13,32] and RAM [18,35], aged rats showed deficits in the rate of acquisition of the task by committing more errors while finding and remembering the platform location. In the present study, latency while learning the initial task was not different between the age groups. However, number of errors were significantly higher in the old rats on all days, making the RAWM an especially sensitive and powerful task to assess age-related deficits in cognitive performance, because studies using the MWM frequently only measure latency, not errors, during acquisition of the task. Additionally, old rats demonstrated cognitive impairment compared to the young controls as evidenced by increased latencies to find the hidden platform on the reversal day when the platform was moved to the opposite quadrant and the rats had to learn a new platform location (and forget the old location). These deficits are very similar to those observed in aged (24 months old) animals, who had difficulty learning a new platform location during reversal training in the MWM [12], and who were impaired in both spatial reference (place discrimination) and working (repeated acquisition) memory measures assessed in the MWM [11]. It is likely that the old rats utilized non-spatial strategies to solve the maze as they showed a lack of spatial preference (less crossings of the old platform location) compared to young animals when the platform was moved on day 4. This result agrees with previous studies which have shown that aged rats, when tested in a probe trial (swim with no platform to measure the use of spatial strategies), spend less time in the training quadrant and have less platform crossings [11,29]. In fact, old animals preferentially use non-spatial
227
strategies to solve tasks when such alternative strategies are available [29]. In a previous study using a different apparatus, we showed that senescent rats have decrements in the ability to build spatial representations of the environment and to utilize this information to detect such changes, even though object recognition is not impaired with age [31]; this lack of spatial response seen in senescent animals is more related to cognitive impairments rather than sensory, motor, or motivational differences [31,33]. Even though old animals swam slower in the present study, differential use of spatial strategies are not due to speed, but rather a decline in the ability to process or retain place (position of a goal with reference to a “map” provided by the configuration of numerous cues in the environment) information [13]. Other studies have shown that old rats swim slower than young controls [12,34], however deficits in latency are not simply attributable to lower swim speeds [12]. Nevertheless, these differences underscore the importance of measuring other parameters that do not depend on swim speed, such as errors which can be easily measured in the RAWM. It is this advantage of the RAWM, the ability to assess both reference and working memory errors simultaneously [15], that produced the largest age-related changes in the present study. Reference memory is believed to reflect learning the trial-independent procedural aspects of the task (spatial cue locations), that is, information that is relevant for many trials, often for the entire experiment [11,20]. Reference memory is consistent from trial to trial, and is required to learn the general rules of any task (e.g., swim to a platform) [11]. In contrast, working memory is trial-dependent and describes the ability of the subject to hold this trial-dependent information (places previously visited) in memory [11,20]. Working memory involves the retention of trial-specific or trial- unique information for short periods of time, and it is necessary to remember both the type of stimulus presented and the time of stimulus presentation [11]. Old rats have previously been shown to have decrements in both parameters in the MWM and the RAM [6,7,16]. In this study, old rats made more errors when solving the RAWM, and in particular more reference memory errors (entries into an arm that did not contain the platform) on all days subsequent to day 1, while more working memory errors (reentries into an arm) were only made on days 2 and 4, not day 3. It appears that the old rats, when only one arm is baited, more quickly learn not to reenter arms, although they still have trouble finding the correct arm. In fact, it appeared from watching the rats’ swim patterns that the old rats paused in the center of the maze before proceeding down an arm (and then often choosing the wrong arm), whereas the young rats performed more quickly and efficiently, that is, they swam directly to the arm with the platform. Moving the platform on day 4 requires the rats to forget the old platform location and relearn a new location, so errors are initially increased in both
228
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
groups; however, the young rats learn the new task quicker than the old rats as evidenced by lower latencies and less errors. It has been shown that aged rats (24 months) display a form of perseverative behavior in the MWM, that is, they swim around the sides of the pool more than the young rats [25]. These thigmotaxic tendencies prevent them from venturing into the vicinity of the pool in which the platform is located [25]; thigmotaxis in the MWM might be attributed to impaired striatal function (motor stereotype) [19]. In the RAWM, this disadvantage is eliminated because the rat cannot swim around the sides of the pool. Furthermore, using a wet version of the radial maze enforces use of rapid allocentric visuospatial information processing more effectively than the dry, in which it may be possible to use several alternative place navigational strategies, for example, odor cues [14]. Therefore, the RAWM is a reliable, additional test to assess age-related spatial learning and memory deficits, combining the advantages of the MWM and the RAM while minimizing the disadvantages. It has shown to be more sensitive in detecting changes in impaired memory ability than the probe trial in the MWM [1]. Also, whereas the MWM provides for the sequential analysis (i.e., two separate trials) of reference (long-term) and working (short-term) memory, the 8-arm RAWM allows the testing of both simultaneously [15]. Furthermore, the RAWM measures acquisition of the task by measuring short- and long-term memory as errors, which has been shown to be a more sensitive parameter to measure age-related deficits compared to latency. Therefore, the RAWM allows an in-depth analysis of both reference and working memory as a function of age. References [1] Arendash GW, King DL, Gordon MN, Morgan D, Hatcher JM, Hope CE, et al. Progressive, age-related behavioral impairments in transgenic mice carrying both mutant amyloid precursor protein and presenilin-1 transgenes. Brain Res 2000;891:42–53. [2] Barnes CA. Aging and the physiology of spatial memory. Neurobiol Aging 1988;9:563–8. [3] Bartus RT, Dean RL, Beer B, Lippa AS. The cholinergic hypothesis of geriatric memory dysfunction. Science 1982;217:408–17. [4] Bimonte HA, Denenberg VH. Sex differences in vicarious trial-anderror behavior during radial arm maze learning. Physiol Behav 2000;68:495–9. [5] Bimonte HA, Hyde LA, Hoplight BJ, Denenberg VH. In two species, females exhibit superior working memory and inferior reference memory on the water radial-arm maze. Physiol Behav 2000;70: 311–7. [6] Brandeis R, Brandys Y, Yehuda S. The use of the Morris water maze in the study of memory and learning. Int J Neurosci 1989;48:29– 69. [7] Brandeis R, Dachir S, Sapir M, Levy A, Fisher A. Reversal of age-related cognitive impairments by an M1 cholinergic agonist, AF102B. Pharmacol Biochem Behav 1990;36:89–95. [8] Buresova O, Bures J, Oitzl MS, Zahalka A. Radial maze in the water tank: an aversively motivated spatial working memory task. Physiol Behav 1985;34:1003–5.
[9] Devan BD, Goad EH, Petri HL. Dissociation of hippocampal and striatal contributions to spatial navigation in the water maze. Neurobiol Learn Mem 1996;66:305–23. [10] Diamond DM, Park CR, Heman KL, Rose GM. Exposing rats to a predator impairs spatial working memory in the radial arm water maze. Hippocampus 1999;9:542–52. [11] Frick KM, Baxter MG, Markowska AL, Olton DS, Price DL. Age-related spatial reference and working memory deficits assessed in the water maze. Neurobiol Aging 1995;16:149–60. [12] Gage FH, Dunnett SB, Bjorklund A. Spatial learning and motor deficits in aged rats. Neurobiol Aging 1984;5:43–8. [13] Gallagher M, Pelleymounter MA. Spatial learning deficits in old rats: a model for memory decline in the aged. Neurobiol Aging 1988;9:549–56. [14] Hodges H. Maze procedures: the radial-arm and water maze compared. Cogn Brain Res 1996;3:167–81. [15] Hyde LA, Hoplight BJ, Denenberg VH. Water version of the radial-arm maze: learning in three inbred strains of mice. Brain Res 1998;785:236–44. [16] Ingram DK, Jucker M, Spangler EL. Behavioral manifestations of aging. In: Mohr U, Cungworth DL, Capen CC, editors. Pathobiology of the aging rat, vol. 2. Washington, DC: ILSI Press; 1994. p. 149–70. [17] Joseph JA. The putative role of free radicals in the loss of neuronal functioning in senescence. Integ Physiol Behav Sci 1992;27:216– 27. [18] Kobayashi S, Kametani H, Ugawa Y, Osanai M. Age difference of response strategy in radial maze performance of Fischer-344 rats. Physiol Behav 1988;42:277–80. [19] Lindner MD, Schallert T. Aging and atropine effects on spatial navigation in the Morris water task. Behav Neurosci 1988;102: 621–4. [20] Luine VN, Richards ST, Wu VY, Beck KD. Estradiol enhances learning and memory in a spatial memory task and effects levels of monoaminergic neurotransmitters. Horm Behav 1998;34:149–62. [21] McDonald RJ, White NM. Parallel information processing in the water maze: evidence for independent memory systems involving dorsal striatum and hippocampus. Behav Neural Biol 1994;61:260– 70. [22] Morgan D, Diamond DM, Gottschall PE, Ugen KE, Dickey C, Hardy J, et al. A peptide vaccination prevents memory loss in an animal model of Alzheimer’s disease. Nature 2000;408:982–5. [23] Morris RGM. Spatial localization does not require the presence of local cues. Learn Motiv 1981;12:239–61. [24] Morris R. Developments of a water-maze procedure for studying spatial learning in the rat. J Neurosci Meth 1984;11:47–60. [25] Muir JL, Fischer W, Bjorklund A. Decline in visual attention and spatial memory in aged rats. Neurobiol Aging 1999;20:605–15. [26] Oliveira MGM, Bueno OFA, Pomarico AC, Gugliano EB. Strategies used by hippocampal- and caudate-putamen-lesioned rats in a learning task. Neurobiol Learn Memory 1997;68:32–41. [27] Olton DS. The radial arm maze as a tool in behavioral pharmacology. Physiol Behav 1987;4:793–7. [28] Olton DS, Samuelson RJ. Remembrance of places passed: spatial memory in rats. J Exp Psychol Anim Behav Proc 1976;2:97–116. [29] Rapp PR, Rosenberg RA, Gallagher M. An evaluation of spatial information processing in aged rats. Behav Neurosci 1987;101:3–12. [30] Shear DA, Dong J, Haik-Creguer KL, Bazzet TJ, Albin RL, Dunbar GL. Chronic administration of quinolinic acid in the rat striatum causes spatial learning deficits in a radial arm water maze task. Exp Neurol 1998;150:305–11. [31] Shukitt-Hale B, Casadesus G, Cantuti-Castelvetri I, Joseph JA. Effect of age on object exploration, habituation, and response to spatial and non-spatial change. Behav Neurosci 2001;115:1059–64. [32] Shukitt-Hale B, Mouzakis G, Joseph JA. Psychomotor and spatial memory performance in aging male Fischer 344 rats. Exp Gerontol 1998;33:615–24.
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229 [33] Soffie M, Buhot MC, Poucet B. Cognitive and noncognitive processes involved in selective object exploration: comparison between young adult and old rats. Physiol Behav 1992;52:1029–35. [34] Van der Staay FJ, de Jonge M. Effects of age on water escape behavior and on repeated acquisition in rats. Behav Neural Biol 1993;60:33–41.
229
[35] Wallace JE, Krauter EE, Campbell BA. Animal models of declining memory in the aged: short-term and spatial memory in the aged rat. J Gerontol 1980;35:355–63. [36] Zyzak DR, Otto T, Eichenbaum H, Gallagher M. Cognitive decline associated with normal aging in rats: a neuropsychological approach. Learn Mem 1995;2:1–16.
Effect of age on the radial arm water maze—a test of spatial learning and memory Barbara Shukitt-Hale∗ , John J. McEwen, Aleksandra Szprengiel, James A. Joseph USDA-ARS, Human Nutrition Research Center on Aging at Tufts University, 711 Washington Street, Boston, MA 02111, USA Received 9 October 2001; received in revised form 3 March 2002; accepted 21 February 2003
Abstract Aged rats show decrements in performance on cognitive tasks that require the use of spatial learning and memory. We used the 8-arm radial water maze (RAWM) to measure spatial learning as a function of age in young (6 months) and old (21 months) male F344 rats. Rats were placed in the RAWM in different start arms with the same goal arm for 3 days (five trials/day); the goal arm was changed on day 4. Old rats demonstrated spatial impairment as evidenced by increased latencies to find the hidden platform on day 4. Old rats made significantly more errors, both reference and working memory errors, than young rats on all days. It is likely that the old rats utilized non-spatial strategies to solve the task, and therefore were impaired in learning a new platform location. The RAWM is a reliable, sensitive, and powerful additional test to assess age-related spatial learning and memory deficits, combining the advantages of the Morris water maze and the radial arm maze while minimizing the disadvantages. © 2003 Elsevier Science Inc. All rights reserved. Keywords: Cognition; Reference memory; Working memory; Short-term memory; Long-term memory; Fischer 344 rat
1. Introduction Aged rats show decrements in performance on cognitive tasks that require the use of spatial learning and memory, that is, the ability to acquire a cognitive representation of location in space and the ability to effectively navigate the environment (for reviews see [2,6,13,16]). Memory alterations appear to occur primarily in secondary memory systems and are reflected in the storage of newly acquired information [3,17]. It is thought that the hippocampus mediates allocentric spatial navigation (i.e., place learning), and that the prefrontal cortex is critical to acquiring the rules that govern performance in particular tasks (i.e., procedural knowledge), while the dorsomedial striatum mediates egocentric spatial orientation (i.e., response and cue learning) [9,21,26,36]. Two well known tests used to assess spatial learning and memory in rats are the Morris water maze (MWM) [23,24] and the radial arm maze (RAM) [27,28]. Both of these tests have distinct advantages and disadvantages in assessing spatial learning and memory in rats. Therefore, in the present study, we used a combination of these two tests, the 8-arm radial water maze (RAWM), to measure spatial learning as a function of age in young ∗
Corresponding author. Tel.: +1-617-556-3118; fax: +1-617-556-3222. E-mail address: [email protected] (B. Shukitt-Hale).
and old rats. This test maintains the advantages of both the MWM and the RAM, while eliminating the disadvantages [8,10,14,15,30]. Several studies have used various versions of the RAWM to measure learning in different strains of mice [15], to assess sex differences in learning [4,5], to investigate the effects of stress on spatial working memory [10], and to assess learning and memory deficits that develop in a transgenic mouse model of Alzheimer’s disease [1,22]. The RAWM combines the variable spatial complexity of the RAM and the efficient learning of the MWM [10]. One distinct advantage of the RAWM is the ability to assess both reference and working memory errors simultaneously [15]. Reference memory is believed to reflect learning the trial-independent procedural aspects of the task (spatial cue locations), while working memory is trial-dependent and describes the ability of the subject to hold this trial-dependent information (places previously visited) in memory [11,20]. Furthermore, in the MWM the rat can use a strategy of swimming an appropriate fixed distance from the wall to find the hidden platform, while in the RAWM the rat is forced to swim in either the central open area or one of the arms, and therefore cannot solve the task by adopting this proximal strategy. The rat can also find the platform in the MWM by bumping into it, while this strategy is not possible in the RAWM without an error being assessed. Swimming around the sides of the pool is also not permitted in the RAWM. In
0197-4580/$ – see front matter © 2003 Elsevier Science Inc. All rights reserved. doi:10.1016/S0197-4580(03)00041-1
224
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
the RAM, food or water deprivation is required, while the RAWM is non-appetitively motivated and therefore testing in it is not confounded by differences in food motivation. Dietary interventions can also be tested in the RAWM since food deprivation is not necessary, and multiple trials can be run per day in the RAWM since satiation is not a factor. The effect of odor cues present in the RAM is eliminated in the RAWM. Also, since the rat is forced to swim in the RAWM, the rat cannot choose whether or not to move, as it can in the RAM, so failure to respond is not a confound. Thus, because of its advantages, we used the RAWM in the present study to measure spatial learning and memory as a function of age.
the platform. Once the rat reached the platform, it remained there for 15 s. At the end of each trial, the rat was towel-dried, returned to its home cage, and 30 min elapsed (during which the remaining rats were tested) before being returned to the maze for its next trial, which used the same platform location and a different start position as the previous trial. Performance on each trial was videotaped and analyzed with image tracking software (HVS Image, Hampton, UK), which provided dependent measures such as latency (s) and speed (cm/s). Also, the order of entry into the maze arms was recorded so that the number of errors could be analyzed; errors were broken into reference (long-term) memory (defined as entering an arm that does not contain the platform) and working (short-term) memory errors (reentries into an arm).
2. Materials and methods 2.3. Statistical analyses 2.1. Animals The subjects consisted of two age groups of male F344 rats (NIA colony, Harlan Sprague Dawley, Indianapolis, IN): young, 6 months (n = 10) and old, 21 months (n = 10), weighing between 300 and 440 g. The rats were individually housed in stainless steel mesh suspended cages, provided food and water ad libitum, and maintained on a 12 h light/dark cycle. All animals were observed daily for clinical signs of disease. The old animals weighed significantly more than the young animals (young: 333.30 ± 6.41 g (mean ± S.E.M.); old: 419.50 ± 6.01 g; F(1, 18) = 96.12, P < 0.001). 2.2. Apparatus The radial arm water maze consisted of a circular galvanized pool (152 cm diameter × 59 cm high) filled to a depth of 32 cm with water (23 ◦ C). Within the pool were stainless steel walls (40 cm high × 43 cm long), arranged to make eight swim paths or arms radiating from an open center area. The pool, walls, and circular escape platform (10 cm diameter × 30 cm high) were all colored black, enabling the platform to be hidden from sight. The maze was placed in a room with the lights dimmed, and there were numerous extramaze cues on the walls and scattered throughout the room that the rat could use to navigate the maze. On days 1–3, the platform was submerged 2 cm below the surface of the water and 4 cm from the end of arm 1, switching to arm 4 on day 4. RAWM testing was performed for 4 days, five trials/day with a 30 min intertrial interval. At the beginning of each trial, the rat was gently immersed in the water, facing the wall, at one of seven start positions. Start locations were randomized, with the qualifications that not the same starting location be used twice in a row and that the rat never be started in the arm with the platform. Each rat was allowed 120 s to escape onto the platform; if the rat failed to escape within this time, it was guided to
For each behavioral measure, repeated measures analysis of variance (ANOVA) models with one between-subjects factor (age) and one within-subjects factor (trials) were performed using Systat (SPSS, Inc., Chicago, IL) to test for statistical significance at the P ≤ 0.05 level. Separate one-way ANOVAs or t-tests were run to determine which trials were different between groups when a significant overall group difference or group × trial interaction was found. 3. Results Old rats demonstrated spatial learning and memory impairment in the RAWM, particularly on error measures and on the reversal day when the platform was moved. On days 1–3, there was significant improvement over time in water maze acquisition (learning) performance (i.e., latency to find the platform), F(4, 72) = 6.69, P < 0.001; improvements in performance were shown by both the old and young groups (Fig. 1). There was no difference in overall acquisition between groups on days 1–3, as latency measures were similar for the two age groups (P > 0.05). It was on day 4, when the location of the platform was changed from arm 1 to arm 4, that the old rats exhibited the most dramatic learning deficits as evidenced by increased latencies to find the hidden platform (subsequent to trial 1) (main effect of age: F(1, 18) = 3.94, P = 0.06; age × trial interaction: F(4, 72) = 3.29, P < 0.05; Fig. 1). When the platform location was moved on day 4, old rats initially found the new platform location more quickly (although not significantly) than the young on trial 1. The young rats, because they were using spatial strategies, took longer to find the new location since they initially searched longer in the former platform position, shown by more crossings of the old platform location (6.1 ± 1.01 crossings of the old platform location for young versus 3.00±0.84 for old rats; F(1, 18) = 5.58, P < 0.05). However, the young rats learned the new task faster than the old group, as the performance of the
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
225
Fig. 1. Escape latencies (s, mean ± S.E.M.) to find the hidden platform in the RAWM for days 1–3 (trials 1–5) and for reversal testing on day 4 (trials 1–5) for young (6 months) and old (21 months) F344 rats.
young improved by the second trial onwards, but the performance of the old group did not. By trials 4 and 5, the latency to find the platform for the young group was significantly less (P < 0.01) than the old group (Fig. 1). Old rats made significantly more errors than young rats on all days (Fig. 2). For days 1–3, ANOVA showed a significant age × trial interaction for total errors, F(14, 252) = 2.27, P < 0.01; although the main effect of age failed to reach statistical significance, F(1, 18) = 3.08, P = 0.09. More specifically, on day 1, young rats initially made more errors on trial 1 but quickly learned the task, F(4, 72) = 3.39, P < 0.05. On day 2, F(1, 18) = 6.13, P < 0.05; and day 3, F(1, 18) = 7.10, P < 0.05; there was a main effect of age, with the old rats making significantly more errors than
the young rats. On day 4, when the platform was moved, again the young rats initially made more errors (although not significantly more) before learning the new platform location, F(4, 72) = 2.58, P < 0.05; the old rats made significantly more errors on trials 2, 4, and 5 (P < 0.05) compared to the young group (Fig. 2). When errors were broken down into reference memory (Fig. 3A) and working memory errors (Fig. 3B), it was found that old rats made significantly more reference memory, F(14, 252) = 2.12, P < 0.05; and working memory, F(14, 252) = 2.18, P < 0.01; errors than young rats on days 1–3, as shown by significant age × trial interactions (again, however, the main effect of age failed to reach significance for reference memory, F(1, 18) = 2.89, P = 0.11;
Fig. 2. Total number of errors made (mean ± S.E.M.) while finding the hidden platform in the RAWM for days 1–3 (trials 1–5) and for reversal testing on day 4 (trials 1–5) for young (6 months) and old (21 months) F344 rats.
226
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
Fig. 3. Total number of reference memory (A) and working memory (B) errors made (mean ± S.E.M.) while finding the hidden platform in the RAWM for days 1–3 (trials 1–5) and for reversal testing on day 4 (trials 1–5) for young (6 months) and old (21 months) F344 rats. Reference memory errors are defined as entering an arm that does not contain the platform, while working memory errors are reentries into an arm.
Fig. 4. Speed (cm/s, mean ± S.E.M.) to find the hidden platform in the RAWM for days 1–3 (trials 1–5) and for reversal testing on day 4 (trials 1–5) for young (6 months) and old (21 months) F344 rats.
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
and working memory, F(1, 18) = 1.97, P = 0.18, errors). Old rats made more reference memory errors compared to young rats on day 2, F(1, 18) = 6.71, P < 0.05; day 3, F(1, 18) = 10.06, P < 0.01; and day 4, F(1, 18) = 9.35, P < 0.01; and more working memory errors on day 2, F(1, 18) = 4.07, P ≤ 0.05; and day 4, F(1, 18) = 6.68, P < 0.05; but not day 3. On day 1, young rats initially made more working memory errors but steadily improved over the trials on this day and did not reenter the arms as frequently as old rats in later trials, F(4, 72) = 3.91, P < 0.01. On days 2–4, young rats swam faster than old animals (day 2, F(1, 18) = 6.32, P < 0.05; day 3, F(1, 18) = 10.22, P < 0.01; day 4, F(1, 18) = 13.29, P < 0.01; Fig. 4). However, there was no difference in speed between the groups on day 1, ruling out gross motor impairment. Instead, the young animals swam faster as the trials progressed, while the old rats maintained one speed or did not increase their speed to the extent of the young, particularly on days 2 and 4 (day 2, F(4, 72) = 3.56, P < 0.05; day 4, F(4, 72) = 4.99, P < 0.01; Fig. 4).
4. Discussion In the RAWM, as in the MWM [13,32] and RAM [18,35], aged rats showed deficits in the rate of acquisition of the task by committing more errors while finding and remembering the platform location. In the present study, latency while learning the initial task was not different between the age groups. However, number of errors were significantly higher in the old rats on all days, making the RAWM an especially sensitive and powerful task to assess age-related deficits in cognitive performance, because studies using the MWM frequently only measure latency, not errors, during acquisition of the task. Additionally, old rats demonstrated cognitive impairment compared to the young controls as evidenced by increased latencies to find the hidden platform on the reversal day when the platform was moved to the opposite quadrant and the rats had to learn a new platform location (and forget the old location). These deficits are very similar to those observed in aged (24 months old) animals, who had difficulty learning a new platform location during reversal training in the MWM [12], and who were impaired in both spatial reference (place discrimination) and working (repeated acquisition) memory measures assessed in the MWM [11]. It is likely that the old rats utilized non-spatial strategies to solve the maze as they showed a lack of spatial preference (less crossings of the old platform location) compared to young animals when the platform was moved on day 4. This result agrees with previous studies which have shown that aged rats, when tested in a probe trial (swim with no platform to measure the use of spatial strategies), spend less time in the training quadrant and have less platform crossings [11,29]. In fact, old animals preferentially use non-spatial
227
strategies to solve tasks when such alternative strategies are available [29]. In a previous study using a different apparatus, we showed that senescent rats have decrements in the ability to build spatial representations of the environment and to utilize this information to detect such changes, even though object recognition is not impaired with age [31]; this lack of spatial response seen in senescent animals is more related to cognitive impairments rather than sensory, motor, or motivational differences [31,33]. Even though old animals swam slower in the present study, differential use of spatial strategies are not due to speed, but rather a decline in the ability to process or retain place (position of a goal with reference to a “map” provided by the configuration of numerous cues in the environment) information [13]. Other studies have shown that old rats swim slower than young controls [12,34], however deficits in latency are not simply attributable to lower swim speeds [12]. Nevertheless, these differences underscore the importance of measuring other parameters that do not depend on swim speed, such as errors which can be easily measured in the RAWM. It is this advantage of the RAWM, the ability to assess both reference and working memory errors simultaneously [15], that produced the largest age-related changes in the present study. Reference memory is believed to reflect learning the trial-independent procedural aspects of the task (spatial cue locations), that is, information that is relevant for many trials, often for the entire experiment [11,20]. Reference memory is consistent from trial to trial, and is required to learn the general rules of any task (e.g., swim to a platform) [11]. In contrast, working memory is trial-dependent and describes the ability of the subject to hold this trial-dependent information (places previously visited) in memory [11,20]. Working memory involves the retention of trial-specific or trial- unique information for short periods of time, and it is necessary to remember both the type of stimulus presented and the time of stimulus presentation [11]. Old rats have previously been shown to have decrements in both parameters in the MWM and the RAM [6,7,16]. In this study, old rats made more errors when solving the RAWM, and in particular more reference memory errors (entries into an arm that did not contain the platform) on all days subsequent to day 1, while more working memory errors (reentries into an arm) were only made on days 2 and 4, not day 3. It appears that the old rats, when only one arm is baited, more quickly learn not to reenter arms, although they still have trouble finding the correct arm. In fact, it appeared from watching the rats’ swim patterns that the old rats paused in the center of the maze before proceeding down an arm (and then often choosing the wrong arm), whereas the young rats performed more quickly and efficiently, that is, they swam directly to the arm with the platform. Moving the platform on day 4 requires the rats to forget the old platform location and relearn a new location, so errors are initially increased in both
228
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229
groups; however, the young rats learn the new task quicker than the old rats as evidenced by lower latencies and less errors. It has been shown that aged rats (24 months) display a form of perseverative behavior in the MWM, that is, they swim around the sides of the pool more than the young rats [25]. These thigmotaxic tendencies prevent them from venturing into the vicinity of the pool in which the platform is located [25]; thigmotaxis in the MWM might be attributed to impaired striatal function (motor stereotype) [19]. In the RAWM, this disadvantage is eliminated because the rat cannot swim around the sides of the pool. Furthermore, using a wet version of the radial maze enforces use of rapid allocentric visuospatial information processing more effectively than the dry, in which it may be possible to use several alternative place navigational strategies, for example, odor cues [14]. Therefore, the RAWM is a reliable, additional test to assess age-related spatial learning and memory deficits, combining the advantages of the MWM and the RAM while minimizing the disadvantages. It has shown to be more sensitive in detecting changes in impaired memory ability than the probe trial in the MWM [1]. Also, whereas the MWM provides for the sequential analysis (i.e., two separate trials) of reference (long-term) and working (short-term) memory, the 8-arm RAWM allows the testing of both simultaneously [15]. Furthermore, the RAWM measures acquisition of the task by measuring short- and long-term memory as errors, which has been shown to be a more sensitive parameter to measure age-related deficits compared to latency. Therefore, the RAWM allows an in-depth analysis of both reference and working memory as a function of age. References [1] Arendash GW, King DL, Gordon MN, Morgan D, Hatcher JM, Hope CE, et al. Progressive, age-related behavioral impairments in transgenic mice carrying both mutant amyloid precursor protein and presenilin-1 transgenes. Brain Res 2000;891:42–53. [2] Barnes CA. Aging and the physiology of spatial memory. Neurobiol Aging 1988;9:563–8. [3] Bartus RT, Dean RL, Beer B, Lippa AS. The cholinergic hypothesis of geriatric memory dysfunction. Science 1982;217:408–17. [4] Bimonte HA, Denenberg VH. Sex differences in vicarious trial-anderror behavior during radial arm maze learning. Physiol Behav 2000;68:495–9. [5] Bimonte HA, Hyde LA, Hoplight BJ, Denenberg VH. In two species, females exhibit superior working memory and inferior reference memory on the water radial-arm maze. Physiol Behav 2000;70: 311–7. [6] Brandeis R, Brandys Y, Yehuda S. The use of the Morris water maze in the study of memory and learning. Int J Neurosci 1989;48:29– 69. [7] Brandeis R, Dachir S, Sapir M, Levy A, Fisher A. Reversal of age-related cognitive impairments by an M1 cholinergic agonist, AF102B. Pharmacol Biochem Behav 1990;36:89–95. [8] Buresova O, Bures J, Oitzl MS, Zahalka A. Radial maze in the water tank: an aversively motivated spatial working memory task. Physiol Behav 1985;34:1003–5.
[9] Devan BD, Goad EH, Petri HL. Dissociation of hippocampal and striatal contributions to spatial navigation in the water maze. Neurobiol Learn Mem 1996;66:305–23. [10] Diamond DM, Park CR, Heman KL, Rose GM. Exposing rats to a predator impairs spatial working memory in the radial arm water maze. Hippocampus 1999;9:542–52. [11] Frick KM, Baxter MG, Markowska AL, Olton DS, Price DL. Age-related spatial reference and working memory deficits assessed in the water maze. Neurobiol Aging 1995;16:149–60. [12] Gage FH, Dunnett SB, Bjorklund A. Spatial learning and motor deficits in aged rats. Neurobiol Aging 1984;5:43–8. [13] Gallagher M, Pelleymounter MA. Spatial learning deficits in old rats: a model for memory decline in the aged. Neurobiol Aging 1988;9:549–56. [14] Hodges H. Maze procedures: the radial-arm and water maze compared. Cogn Brain Res 1996;3:167–81. [15] Hyde LA, Hoplight BJ, Denenberg VH. Water version of the radial-arm maze: learning in three inbred strains of mice. Brain Res 1998;785:236–44. [16] Ingram DK, Jucker M, Spangler EL. Behavioral manifestations of aging. In: Mohr U, Cungworth DL, Capen CC, editors. Pathobiology of the aging rat, vol. 2. Washington, DC: ILSI Press; 1994. p. 149–70. [17] Joseph JA. The putative role of free radicals in the loss of neuronal functioning in senescence. Integ Physiol Behav Sci 1992;27:216– 27. [18] Kobayashi S, Kametani H, Ugawa Y, Osanai M. Age difference of response strategy in radial maze performance of Fischer-344 rats. Physiol Behav 1988;42:277–80. [19] Lindner MD, Schallert T. Aging and atropine effects on spatial navigation in the Morris water task. Behav Neurosci 1988;102: 621–4. [20] Luine VN, Richards ST, Wu VY, Beck KD. Estradiol enhances learning and memory in a spatial memory task and effects levels of monoaminergic neurotransmitters. Horm Behav 1998;34:149–62. [21] McDonald RJ, White NM. Parallel information processing in the water maze: evidence for independent memory systems involving dorsal striatum and hippocampus. Behav Neural Biol 1994;61:260– 70. [22] Morgan D, Diamond DM, Gottschall PE, Ugen KE, Dickey C, Hardy J, et al. A peptide vaccination prevents memory loss in an animal model of Alzheimer’s disease. Nature 2000;408:982–5. [23] Morris RGM. Spatial localization does not require the presence of local cues. Learn Motiv 1981;12:239–61. [24] Morris R. Developments of a water-maze procedure for studying spatial learning in the rat. J Neurosci Meth 1984;11:47–60. [25] Muir JL, Fischer W, Bjorklund A. Decline in visual attention and spatial memory in aged rats. Neurobiol Aging 1999;20:605–15. [26] Oliveira MGM, Bueno OFA, Pomarico AC, Gugliano EB. Strategies used by hippocampal- and caudate-putamen-lesioned rats in a learning task. Neurobiol Learn Memory 1997;68:32–41. [27] Olton DS. The radial arm maze as a tool in behavioral pharmacology. Physiol Behav 1987;4:793–7. [28] Olton DS, Samuelson RJ. Remembrance of places passed: spatial memory in rats. J Exp Psychol Anim Behav Proc 1976;2:97–116. [29] Rapp PR, Rosenberg RA, Gallagher M. An evaluation of spatial information processing in aged rats. Behav Neurosci 1987;101:3–12. [30] Shear DA, Dong J, Haik-Creguer KL, Bazzet TJ, Albin RL, Dunbar GL. Chronic administration of quinolinic acid in the rat striatum causes spatial learning deficits in a radial arm water maze task. Exp Neurol 1998;150:305–11. [31] Shukitt-Hale B, Casadesus G, Cantuti-Castelvetri I, Joseph JA. Effect of age on object exploration, habituation, and response to spatial and non-spatial change. Behav Neurosci 2001;115:1059–64. [32] Shukitt-Hale B, Mouzakis G, Joseph JA. Psychomotor and spatial memory performance in aging male Fischer 344 rats. Exp Gerontol 1998;33:615–24.
B. Shukitt-Hale et al. / Neurobiology of Aging 25 (2004) 223–229 [33] Soffie M, Buhot MC, Poucet B. Cognitive and noncognitive processes involved in selective object exploration: comparison between young adult and old rats. Physiol Behav 1992;52:1029–35. [34] Van der Staay FJ, de Jonge M. Effects of age on water escape behavior and on repeated acquisition in rats. Behav Neural Biol 1993;60:33–41.
229
[35] Wallace JE, Krauter EE, Campbell BA. Animal models of declining memory in the aged: short-term and spatial memory in the aged rat. J Gerontol 1980;35:355–63. [36] Zyzak DR, Otto T, Eichenbaum H, Gallagher M. Cognitive decline associated with normal aging in rats: a neuropsychological approach. Learn Mem 1995;2:1–16.