Behavioral Neuroscience 2013, Vol. 127, No. 6. 867-877
© 2013 American Psychological Association 0735-7044/13/$ ! 2.00 DOI: 10.1037/a0034931
The Effects of Bilateral Lesions to the Dorsal Tegmental Nucleus on Spatial Learning in Rats Jessica A. Dwyer, Matthew L. Ingram, Anna C. Snow, Christina M. Thorpe, Gerard M. Martin, and Darlene M. Skinner Memorial University of Newfoundland The head-direction (HD) signal is believed to originate in the dorsal tegmental nucleus (DTN) and lesions to this structure have been shown to disrupt HD cell firing in other areas along the HD cell circuit. To investigate the role of the DTN in spatial navigation, rats with bilateral, electrolytic (Experiment 1), or neurotoxic (Experiment 2) lesions to the DTN were compared with sham controls on two tasks that differed in difficulty and could be solved using directional heading. Rats were first trained on a direction problem in a water T maze where they learned to travel either cast or west from two locations in the experimental room. DTN-lesioned rats were impaired relative to sham controls, both early in training, on the first block of eight trials, and on the total trials taken to reach criterion. In the food-foraging task, rats were trained to leave a home cage at the periphery of a circular table, find food in the center of the table and return to the home cage. Again, DTN-lesioned rats were impaired relative to sham rats, making more errors on the return component of the foraging trip. These data extend previous cell-recording studies and behavioral tests in which rats with electrolytic DTN lesions were used, and they demonstrate the importance of the direction system to spatial leaming. Keywords: direction learning, head-direction cells, water maze, T maze, foraging task
When an animal navigates its environrnent, it relies on knowledge of the direction in which it is headed and its relative location in space (Taube, 2007). These two components of navigation are well-represented in a spatial network within the mammalian brain that contains place cells, grid cells, and head-direction (HD) cells (Fyhn, Molden, Witter, Moser, & Moser, 2004; Hafting, Fyhn, Molden, Moser, & Moser, 2005; Moser, Kropff, & Moser, 2008; Muller, 1996; O'Keefe & Dostrovsky, 1971; Taube, 1998; Taube, 2007; Taube, Muller, & Ranck, 1990a, 1990b). O'Keefe and Dostrovsky's (1971) initial report of neurons in the rat hippocampus that fired in a location-specific manner led to the suggestion that the hippocampus was crucial to the formation of spatial maps (O'Keefe & Nadel, 1978). More recent studies have revealed that place cells are part of a broader network for location that also includes grid cells of the entorhinal cortex (Fyhn et al., 2004; Hafting et al., 2005). Like hippocampal place cells, grid cells show spatially selective firing, but unlike place cells, they have multiple firing fields that form a periodic triangular array over the entire environment explored by the animal. Although place cells and grid cells provide a means of knowing our positions in the environment, to get from place to place probably requires heading information fiom HD cells (Kubie & Fenton, 2009; Moser et al., 2008; Taube,
1998, 2007). HD cells fire when an animal's head is pointed in a particular direction and, unlike place cells or grid cells, are independent of the animal's location (Taube, 2007: Taube et al., 1990a, 1990b). The preferred firing direction of HD cells can be controlled by both external (allothetic) cues (e.g., visual) and intemal (idiothetic) cues (e.g., vestibular and proprioceptive; Taube, 2007; Taube & Burton, 1995; Taube et al., 1990b). Although HD cells have been identified in various brain regions (Chen, Lin, Green, Barnes, & McNaughton, 1994; Cho & Sharp, 2001; Mizumori, Ragozzino, & Cooper, 2000; Taube, 1995; Taube et al., 1990a: Wiener. 1993; see Wiener & Taube, 2005 for review), they appear to be most abundant, and best studied, in limbic areas that form part of the classical Papez circuit (see Taube, 2007 for review). Through a combination of lesion and cell-recording studies, it has been determined that the HD signal probably originates in the reciprocal connections between the dorsal tegmental nucleus (DTN) and the lateral mammillary nuclei (LMN), and it is processed through mostly serial projections (with some reciprocal connections) from DTN^LMN^anterior dorsal nucleus of the thalamus (ADN)^postsubiculuin/parasubiculum (PoS/PaS) and retrospenial cortex (RSP; see Clark & Taube, 2012 for review). The connections between the PoS and the entorhinal cortex, the major input to the hippocampus, permit the directional signal to be integrated with location information from grid cells and place cells (see Taube, 2007 for review). Furthermore, developmental studies have revealed that HD cells exhibit adult-like firing properties earlier than place and grid cells (Langston et al., 2010; Wills, Cacucci, Burgess, & O'Keefe, 2010), leading to speculation that they may set up the entire spatial network in the mammalian brain. Although HD cells are thought to underlie an animal's sense of direction (Taube, 1998, 2007), correlations between HD cell ac-
Jessica A. Dwyer, Matthew L. Ingram, Anna C. Snow, Christina M. Thorpe. Gerard M. Martin, and Darlene M. Skinner. Department of Psychology, Memorial University of Newfoundland, St. John's, Newfoundland, Canada. This work was supported by an NSERC grant to Darlene M. Skinner. Correspondence concerning this article should be addressed to Darlene M. Skinner, Department of Psychology, Memorial University of Newfoundland, St. John's, NL AlB 3X9, Canada. E-mail: [email protected] 867
868
DWYER, INGRAM, SNOW, THORPE, MARTIN, AND SKINNER
tivity and behavior are often weak (Muir & Taube, 2002, 2004; Stackman, 2011) and, compared with the vast literature on the effects of damage to the hippocampal system on spatial navigation, relatively few lesion studies have been conducted to examine the behavioral effects of damage to the HD system. Indeed, as suggested by Stackman (2011), the lack of correlation between HDcell activity and spatial performance may be due in part to recording cell activity in brain areas that are not engaged by the behavioral task employed. He suggested an empirical approach in which a researcher first finds a behavioral task that requires an animal to use directional information, and then assess whether performance on the task depends on the activity of HD cells. Behavioral work in our lab has demonstrated a key role for direction (or orientation) to the solution of a variety of spatial problems. Through numerous experiments designed to simplify tasks so that the precise role of place, response, and direction strategies, as well as the contribution of apparatus cues versus distal cues, could be determined, we have repeatedly shown that starting orientation is important to the solution of many spatial tasks (Skinner et al., 2003; Skinner, Horne, Murphy, & Martin, 2010; Whyte, Martin, & Skinner, 2009). Using a paradigm in which the location of the apparatus changes across trials, we have shown that changes in orientation at the start points of the different maze positions are critical to the solution of place and direction problems (Peckford, McRae, Thorpe, Martin, & Skinner, 2013; Skinner et al., 2010) and facilitate response-reversal learning (Wright, Williams, Evans, Skinner, & Martin, 2009). The discrimination between starting orientations is not dependent on the view of distal visual cues from the maze, but does require some access to room cues (Peckford et al., 2013). In that study, rats were trained from two maze positions to swim in a consistent direction from choice point to goal. When the maze was later moved to two new locations, rats required to make the same response based on start-arm orientation showed no disruption in performance. The same pattem of results was obtained if the room was made dark prior to the start of each trial. However, if the rats were brought into an already darkened room then performance was disrupted. Given that HD cells are thought to underlie an animal's sense of direction, and we have developed tasks in which sense of direction or orientation is critical to the solution, we sought to determine if lesions to a component of the HD cell pathway would disrupt our behavioral measures of direction leaming. The purpose of the present study was to assess the behavioral effects of lesions to the DTN on spatial navigation. Because lesions to the DTN have been shown to disrupt the HD cell signal in other areas of the HD cell circuit, including the LMN (Bassett, Tullman, & Taube, 2007), ADN (Bassett et al., 2007), and PoS (see Taube, 2007 for review), and there is speculation that the HD signal originates subcortically in the reciprocal connection between the DTN and the LMN (Clark & Taube, 2012), we predicted that bilateral lesions to the DTN would impair performance on tasks in which directional heading is important. Very few studies have been conducted that assess the behavioral effects of DTN lesions. Although the DTN has HD cells and neurons modulated by both head direction and angular head velocity, the vast majority of the cells within the DTN fire solely in relation to the animal's angular head velocity (Taube & Bassett, 2003). Frohardt, Bassett, and Taube (2006) showed that rats with lesions to the DTN were severely impaired in a food-carrying task.
whereas rats with ADN lesions were only mildly impaired, despite the fact that the ADN has a much higher proportion of HD cells than the DTN. These results imply that some compensatory mechanism is available following damage to the ADN that is not available after DTN lesions and are consistent with the notion mentioned above that the DTN is important in generating the HD signal. A recent study by Clark et al. (2013) showed that lesions to the DTN impaired navigation based on distal landmarks using a modified water-maze procedure (see Hamilton et al, 2008). In both of these earlier studies, the rats were given electrolytic lesions to the DTN. In the present study, we assessed the effects of both electrolytic and neurotoxic lesions to the DTN.
Experiment 1 In the first experiment, rats with electrolytic lesions to the DTN were compared with sham controls on two tasks in which directional heading is important. In the first task, the two groups of rats were trained on a direction problem in a water T maze where they learned to travel either east or west from two start locations to an escape platform (Whyte et al., 2009). Previous work has shown that directional heading in the start arm, not cues at the goal location or the distal visual cues at the start location, is critical to the solution of this direction task (Peckford et al., 2013). Because other strategies such as place and response learning have been ruled out and only two maze positions are used, this task is simpler than other "pure" direction problems. For example, in the Clark et al. (2013) study, acquisition of direction learning was assessed in a task in which four different maze positions were used. Upon completion of the water T-maze task, the same rats were then trained on a food-foraging task (Whishaw & Maaswinkel, 1998), similar to that used by Frohardt et al. (2006) in which they were required to leave a home hole, find food, and then return home. Given the behavioral effects of electrolytic DTN lesions previously reported by Frohardt et al. (2006), and more recently by Clark et al. (2013), we predicted that lesions to the DTN would impair performance on both of these tasks.
Method Subjects. Twenty-four male Long-Evans rats, obtained from Charles River Company (St. Constant, Quebec, Canada) and weighing 250-300 g at the start of surgery, were used. The rats were singly housed in transparent plastic cages (45 X 25 X 21 cm) with metal lids in a temperature-controlled colony room (20 °C) and maintained on a 12-hr light/dark cycle with lights on at 7:00 a.m. All rats had continuous access to food and water in their horne cages during water T-maze training. One week prior to the foodforaging task, the rats were placed on a food-deprivation schedule to maintain their weights at 85% of their free-feeding weights. All procedures were approved by Memorial University of Newfoundland's Institutional Committee on Animal Care and followed the Canadian Council on Animal Care guidelines. Surgery. Rats were anesthetized with an intraperitoneal injection of 0.24M chloral hydrate (1 mLlWOg) and placed in a stereotaxic instrument (Model 900, Kopf, Tujunga, CA) in the skull-flat position. Additional supplements of 0.5 ml of chloral hydrate were administered as necessary. After a scalp incision was made, holes were drilled in the skull (one per hemisphere) using the bregma as
DORSAL TEGMENTAL LESIONS AND SPATIAL LEARNING
a reference point (anterior/posterior (AP); —11.7, medial/lateral (ML): ± 0.3). Sixteen rats were given lesions aimed at both the dorsal and ventral subdivisions of the DTN. A stainless steel electrode (NE-100, Rhodes Medical Instrument, Summerland, CA) was lowered 7.0 mm from the surface of the brain at a 20° angle, and 0.30 milliamps current was applied for 15 s in each hemisphere. Sham rats (n = 8) underwent the same procedure, but no current was applied. All rats were given a minimum of one week to recover before any behavioral testing. Apparatus and materials. The water T maze consisted of a Plexiglas plus maze inserted into a circular metal tank (120 cm in diameter, 31 cm high). Plexiglas walls surrounded the perimeter of the tank and both the walls and the plus maze extended 31 cm above the metal tank. The arms of the plus maze were 11.5 cm wide and 52.5 cm long. The entire apparatus was placed on a metal frame with wheels to facilitate location changes. The plus maze was converted to a T maze using a clear Plexiglas wall held in place by butterfly clips (See Figure 1). The Plexiglas wall prevented access to the arm opposite the start arm, but it did not obstruct the view of the entire tank and the surrounding room cues. The water level was kept approximately 3 cm below the top of the metal tank (28 cm deep). The water was left overnight to equilibrate with the room temperature (approximately 20 °C) and was made opaque by adding approximately 250 ml of nontoxic white Tempera paint (Rich Art Color Company, Northvale, NJ). The escape platform (11.5 cm in diameter and 26.5 cm high) was constructed from white polyvinyl chloride pipe filled with sand and attached to a Plexiglas base for stability. The escape platform was hidden from view approximately 1.5 cm below the surface of the opaque water. Water-maze training took place in a room (528 X 464 X 267 cm) located on the ground floor that had windows covering the north wall and doors located in the center of both the south and east walls. In the southwest corner of the room there was a sink. Shelves lined the west wall and half of the east wall. The southeast comer of the room contained a coat rack and a garbage can. The northwest corner had a small desk at which the experimenter sat to record data. Animal cages were arranged along a table below the shelving on the west wall. For the food-foraging task, the open field maze consisted of a wooden, white circular table (204 cm in diameter) raised 75 cm above the floor. Eight holes (11.5 cm in diameter) were evenly distributed (45° apart and centered 13.5 cm from the edge of the table) around the perimeter of the table and three white, square
West
East
Figure I. A schematic diagram of the water T-maze positions. Half of the rats were trained to go west from maze positions B and D and the other half were trained to go east from maze positions A and C. Anows indicate the correct path and solid black lines indicate a Plexiglas barrier. The entire apparatus (circular tank plus T maze) was translocated between trials.
wooden food cups ( 6 X 6 X 4 cm) were placed equidistant (50 cm) from each other in a triangular configuration at the center (See Figure 2). Large food pellets (1 g; BioServ, Frenchtown, NJ) were used to reinforce the foraging behavior. During training, a pellet was placed in one of the three food cups. The rats were held in wire-mesh cages (20 X 25 X 19 cm) for transport to the training room. The wire cages could be attached to runners beneath the holes at the periphery of the circular table. Wooden stairs were placed in the wire cage with the rat at the beginning of each trial to allow for easy access to the table. Once the wire cage was placed beneath a hole, a rat could leave the cage by climbing up the stairs and on to the table and return to the cage by climbing back down the stairs. For pretraining, the circular table was situated in the center of a ground-floor training room (680 X 507 X 267 cm) that had doors on the east and west walls and a small window on the west wall. There were cupboards, counters, and boxes along all four walls. In the northwest comer, there was a desk at which an experimenter sat to record data. The rack (150 X 50 X 165 cm), used to transport all of the rats in their wire-mesh cages, was pushed near the door along the east wall. For training on the food-foraging task, the circular table was situated in the center of a different training room (478 X 338 X 253 cm) located in the basement that had a door on the north wall. There were boxes along the west wall, a table and stool in the northwest comer of the room, and a sink in the southwest comer. In the northeast corner, there was a desk at which an experimenter sat to record data. The transport rack was pushed against the west wall. Procedure. Water T-maze training. The 24 rats (16 lesioned, 8 shams) were trained to swim in a consistent direction (half went east, half went west) to locate the hidden escape platform from two different maze positions (see Figure 1). The order of maze positions varied daily with no more than two trials in a row from the same position. Rats were brought into the training room with the water maze in three squads of eight and placed individually in transparent plastic holding cages (45 X 25 X 21 cm) that were lined with absorbent paper towel, and airanged on the table along the west wall. The remaining rats were housed in their home cages on metal racks in the hallway outside the training room. For each trial, the rats were carried in their holding cages around the water maze in a counterclockwise direction to a chair positioned at the start arm. A trial began when one experimenter (standing at the start arm for the entire trial) lowered a rat to water level in the start arm facing the Plexiglas wall of the maze. A second experimenter sitting at the desk in the northwest corner of the room recorded the arms visited by the rat and the latency (in seconds) to locate the hidden escape platform. A rat was considered to have made an arm choice when its entire body, not including its tail, was inside an arm. A correct trial was one in which the rat entered the arm containing the escape platform and climbed onto the platform, without previously visiting any other arms. After sitting on the platform for 5 s, the rat was returned to its holding cage at the start arm and carried back (in a clockwise direction) to the table along the west wall. If a rat did not locate the platform within 60 s, it was placed on the platform by the experimenter. The rats in each squad were run sequentially with an intertriai interval of approximately 10 min until each had received its eight daily trials. The rats were given eight trials per day until a criterion of 18/20 correct trials was
870
DWYER, INGRAM, SNOW, THORPE. MARTIN, AND SKINNER
Hole Food cup
Figure 2. A schematic diagram of the food-foraging apparatus. The black circles represent holes in the wooden table under which a wire-mesh cage was positioned. Only Holes 1, 3, and 5 were used as variable start points. Rats were trained to locate a large food pellet placed in one of three food cups (indicated in gray) and return to the variable home location.
reached. For those DTN-lesioned rats that did not reach this cdterion. training was stopped after 88 trials (double the number of trials taken to reach criterion by the slowest sham rat). Food-foraging task. The rats were first pretrained to leave a hole at the periphery of the table and retrieve a food pellet. To coax tbe rats from their cages, 1-g pellets were initially placed near the opening of the hole and incrementally moved further away with each exposure. Eventually, a single pellet was placed centrally on the table in a wooden cup, as in the foraging task, requiring the rats to leave the start hole and travel to the cup to obtain the food. Rats were given three such exposures per day for 2 min each until they were consistently leaving the home location to obtain food from the cup. The maximum number of trials required for any rat to become comfortable performing the task consistently was 21 (7 days). Once pretraining was completed, the apparatus was moved to a new room and all rats were trained to leave Holes 1,3, and 5 (three trials per day, one from each starting point) and find a pellet that was located in one of the three centrally positioned food cups (see Eigure 2). For each trial, the rats were caixied in their holding cages around the circular maze in a counterclockwise direction and placed below the designated start hole. A trial began when the rat emerged (all four limbs) from the wire-mesh cage, and ended when the rat returned to the cage. The wire-mesh cage was then removed from the table and the rat was returned to the metal transport rack along the west wall. The rats were run sequentially in three squads of eight with an intertdal interval of approximately 20 min. The circular table was rotated 45° every trial and washed with Sunlight soap and water to minimize scent trails. A trial was considered to be correct when the rat retrieved the pellet and immediately returned to the hole from which it started without visiting any other holes. First-choice errors were divided into three types. "Adjacenthole errors" occurred when a rat nose-poked the hole 45° to the left or right of its start hole, "memory errors" consisted of the exploration of Start Holes 1, 3, or 5 from a previous trial, and "other
errors" included visitation to any other hole. Because the rats often made more than one error per trial, we also analyzed total errors by the two groups. Each day, every rat received a tdal starting from Holes 1, 3, and 5 (in random order) going to Food Cup I, 2, or 3 (also randomized). Rats were run until they obtained a pellet from the centrally located food cups 60 times. Tdals in which the rat did not leave its cage or did not find food were not included in the analyses. Histology. Once the behavioral testing was completed, the rats were euthanized with intraperitoneal injections of 15% urethane. The brains were extracted and immediately fixed with cooled 2-methylbutane and stored at - 7 0 °C. Sections were taken at 30 (Jim and stained with cresyl violet Nissl for microscopic (14X) vedfication of the lesion sites, using the rat-brain atlas by Paxinos and Watson (1998) as a guide. To quantify the extent of tissue damage created by the electrolytic lesions, digital pictures were taken at the level of the central and pericentral subregions of the DTN (-9.3 mm postedor to bregma). We outlined and measured undamaged DTN with intact neurons and few glial cells using Image-J software (http://rsbweb.nih.gov/ij/), blind to the animal's individual performance in the behavioral testing. The total amount of tissue damage was calculated using the following equation as per Clark et al. (2013): tissue damage = average area in pixels^ of DTN in control rats — total area in pixels'^ of spared DTN tissue in lesioned rats / average area in pixels" of DTN in control rats X 100%.
Results Figure 3 diplays representative brain sections from a sham control and a DTN-lesioned rat showing both the central and pericentral subregions of the DTN, and a schematic diagram indicating the smallest and largest lesions. Based on histological analysis performed blind to the behavioral performance of the subjects, 13 of the 16 lesioned animals sustained significant damage to the DTN (>70%). Three rats were excluded from the behavioral analysis because of insufficient damage to the DTN and/or extensive damage beyond the boundaries of the DTN. Of the remaining rats, the absence of brain tissue (visible hole) and evidence of tissue damage (darker stain around lesion site and shrinkage of surrounding structures) in the area of the DTN allowed for lesion verification. Eight rats had damage that extended outside the DTN, including damage to the laterodorsal nucleus, the dorsal raphe nucleus and the central gray area. However, as there was no correlation between lesion size and trials to criterion on the water-maze task (r = —0.26, p = .40) we included these rats in the behavioral analyses. Along with the three rats excluded from all of the behavioral analyses due to damage outside the DTN, one additional lesioned rat was excluded from the food-foraging analyses, as it refused to leave the home cage dudng training. Nine of the lesioned rats that completed both behavioral tasks had complete, or near complete, damage of the DTN (>85%), and most important, had complete damage of the central subregion of the DTN. Water T maze. DTN-lesioned rats were impaired eady in training on the direction task. DTN-lesioned rats performed significantly worse than sham controls on the first block of eight trials, i(19) = 5.45, p < .05; see Eigure 4A. The difference between groups continued throughout training as lesioned rats
DORSAL TEGMENTAL LESIONS AND SPATIAL LEARNING
871
•10.04 ram
D
Figure 3. Panels A (posterior) and C (anterior) show images of a sham brain and panels B (posterior) and D (anterior) show corresponding images from a typical lesioned brain with bilateral electrolytic damage (evidenced by large hole) in the area of the DTN. The far right panel shows the anterior and posterior extent of the DTN lesions with hatch marks indicating the largest lesion and the solid shading showing the smallest lesions. The area of the DTN is outlined in white (adapted with permission from The Rat Brain in Stereotaxic Coordinates (4th ed.), by G. Paxinos and C. Watson, 1998, San Diego, California: Academic Press. Copyright 1998 by Elsevier Academic Press).
required sigtilficantly tnore trials than sham controls to reach criterion, ?(19) = 4.60, p < .05; see Figure 4B. Indeed, four of the lesioned rats failed to reach criterion on the task in twice as many trials as the slowest sham control rat. However, given that nine of the lesioned rats did eventually meet criterion, some strategy must have been available that does not depend on the integrity of the DTN. Food-foraging task. Along with the three rats excluded from all of the behavioral analyses due to damage outside the DTN, one additional lesioned rat was excluded from the food-foraging analyses, as it refused to leave the home eage during training. The data from the 60 trials in which rats found food were divided into five blocks of 12 trials for analyses. The DTN-lesioned rats showed severe impairments on the foraging task. A two-way (Group X Trial Block) ANOVA on trials correct revealed a significant effect of group, F(\, 18) = 204.7, p < .05. A two-way (Group X Trial Block) ANOVA on total errors over blocks of 12 trials also revealed a significant effect of group, F(l, 18) = 70.12, p < .05. These results confirm that lesioned rats performed much worse than sham rats (see Figure 5 A and Figure 5B). The absence of a trial-block effect suggests no improvement across trials on either of these measures. The control rats maintained a high level of performance and the DTN-lesioned rats were consistently impaired. A separate two-way (Group X
Error Type) ANOVA examining the different types of etTors made by rats in the two groups on their first hole choice revealed a significant effect of group, F(l, 18) = 204.7, p < .05 and a significant effect of error type, F(2, 36) = 13.74, p < .05. BonfetToni posttests revealed that subjects made more memory errors than adjacent {p < .05) or other (p < .01) errors, which did not differ (p > .05; see Figure 5C).
Experiment 2 The results from the first experiment add to a small but growing literature on the effects of damage to the HD system on spatial learning (e.g., Bett, Wood, & Dudchenko, 2012; Clark et al., 2013; Frohardt et al, 2006; Vann, 2005, 2011). Electrolytic lesions to the DTN produced impairments on a simple direction task in a water T maze. The advantages of this task were that only two maze positions were used, rather than four (as in Clark et al., 2013), and the confined trajectories make choices unambiguous. Although lesioned rats did show improvement on the water-maze task (i.e., reached criterion), they were severely impaired on the foraging task and did not appear to improve. These data are consistent with previous cell-recording studies and behavioral tests with DTNlesioned rats (Clark et al., 2013; Frohardt et al, 2006; see Taube, 2007 for review). In Experiment 2 we assessed whether neurotoxic
DWYER, INGRAM. SNOW. THORPE, MARTIN, AND SKINNER
872
u 4) Ü
m .2 c m 0)
DTN
Sham
Group
the same procedure but no injections were made. All rats were given a minimum of one week to recover before any behavioral testing. Apparatus and materials. The water-maze and foraging apparatus were the same as used in Experiment 1. Water-maze training was conducted in the same room as in Experiment 1. The room used for pretraining on the foraging task (523 X 439 X 267 cm) had windows on the north wall with a desk and a table and computers. There were three doors in the room, one on the west wall, one on the east wall, and one on the south wall. Shelves containing lab supplies lined the east, south and half of the west wall. Undemeath the shelving, in the southeast corner of the room were a countertop and a sink, and a table was positioned along the
12-
-©- Sham - • - DTN
108-
o s
6-
IS
o Sham
420'
DTN
Group Figure 4. (A) Mean (+ SEM) trials correct over the first eight trials for DTN-lesioned rats and sham controls on a direction task in a water T maze. (B) Mean (+ SEM) trials to criterion of DTN-lesioned rats and sham controls on the direction task.
lesions to the DTN would produce similar deficits on these two tasks. We also included a simpler, fixed-hole version of the foodforaging task. Subjects. Sixteen male Long-Evans rats, obtained from Charles River Company (St. Constant, Quebec, Canada) and weighing 250-300 g at the start of surgery, were used. The rats were housed and maintained as in Experiment 1. All procedures were approved by Memorial University of Newfoundland's Institutional Committee on Animal Care and followed the Canadian Council on Animal Care guidelines. Surgery. Rats were anesthetized with an intraperitoneal injection of 0.24 M chloral hydrate (1 mL/100 g) and placed in a stereotaxic instrument (Model 900, Kopf, Tujunga, CA) in the skull-flat position. Additional supplements of 0.5 ml of chloral hydrate were administered as necessary. After a scalp incision was made, holes were drilled in the skull (one per hemisphere) using bregma as a reference point (AP: — 11.7, ML: ± 0.3). Fight rats were given lesions aimed at the DTN and eight rats served as sham controls. Neurotoxic lesions were produced by injecting 0.3 |jil of 100 mM N-methyl-D-aspartate (Sigma Chemical, St. Louis, MO), mixed in phosphate-buffered saline, into each hemisphere using a 1 p,l Hamilton syringe and micro injection unit (Model 5000, Kopf). The syringe was lowered 7.1 mm from the surface of the brain at a 20° angle. The N-methyl-D-aspartate was injected at a constant rate over 6 min, and the needle was left in position for an additional 5 min after the injection. Sham rats (n = 8) underwent
2 3 4 Blocks Of 12 Trials
B 50-
-©- Sham - • - DTN
402 o
30-
1 ^°' 1—
102
0
3
4
Blocks Of 12 Trials
• • Memory Errors • Adjacent Errors ra Other Errors
Sham
DTN Group
Figure 5. (A) Mean ( ± SEM) trials correct over five blocks of 12 trials for DTN-lesioned rats and sham controls on the food-foraging task. (B) Mean(± 5£Af) total number of errors made over five blocks of 12 trials for DTN-lesioned rats and sham controls on the food-foraging task. (C) Mean (-f- SEM) number of first-choice memory, adjacent and other errors made over 60 trials for DTN-lesioned rats and sham controls on the foodforaging task.
DORSAL TEGMENTAL LESIONS AND SPATIAL LEARNING
west wall. The cart housing the animal cages was positioned in a space near the southwest comer of the room; the experimenter sat at a table in the northeast comer. The training room (589 X 465 X 267 cm) for the fixed-hole version of the foraging task was a large room that had been divided into two training areas with a blue curtain in the center of the room. There were two doors in this room, one on the west wall and one on the north wall. The south wall was lined with windows and the west wall was lined with shelves. There were two tables in the room; one along the west wall and one along the east wall. A long countertop (with a shelf above it) ran along the southemmost half of the east wall, and various lab supplies could be found underneath this counter. In the southwest corner of the room, there was a sink. This fixed hole task was completed in the southern half of the room and the experimenter sat by the sink in the northeast corner. The cart housing the animal cages was positioned in a space near the curtain in the center of the room. The training room for the variable-hole version of the foraging task (259 X 589 X 267 cm) had two doors, one on the north wall and one on the east wall. On the west wall was a large air conditioner. In the northwest comer of the room there was a sink and, during experimental trials, the experimenter sat by this sink. Windows lined the south wall and a computer table (with a computer) was positioned in front of these windows. The cart housing the animal cages was positioned by the door on the north wall. Procedure Water T-maze training. The rats were trained to swim in a consistent direction (half went east, half went west) to locate the hidden escape platform from two different maze positions. All details of the training procedure for this phase were identical to those outlined in Experiment 1. Training was stopped at 120 trials for those rats that did not reach criterion. Food-foraging task. The rats were first pretrained to leave a hole at the periphery of the table and retrieve a food pellet, as in Experiment 1. Because the amount of pretraining required was variable, and some rats completed pretraining days before others.
873
all rats were administered refresher trials on the day before regular training began. Upon completion of pretraining, the apparatus was moved to a new room and the rats were trained to retum to a single, fixed hole on the apparatus after foraging for a food pellet (located in one of the three food cups). During training, each rat was pseudorandomly assigned a home hole (1, 3, 5, or 7, as shown in Figure 2). The rats were given three trials/day until they reached a criterion of 9/10 trials correct. Training was stopped at 34 trials for rats in the lesioned group that did not meet criterion; this was double the number of trials required by the slowest sham rat. All other details of the training phase were identical to those of Experiment 1. Upon completion of the fixed-hole version of the task, the apparatus was moved again and the rats were trained on the variable-hole version of the task. All details of the training procedure for this task were identical to those of Experiment 1. Histology. Once the behavioral testing was completed, the rats were euthanized and the brains were processed as in Experiment 1.
Results Two rats (one sham and one lesioned rat) were euthanized due to excessive scratching around the surgical site; the sham rat had completed training in the water maze but the lesioned rat had not. A second lesioned rat died during the fixed-hole version of the foraging task. Figure 6 shows representative brain sections from a sham control and a DTN-lesioned rat. Based on histological analysis performed as per Experiment 1, six of the seven lesioned animals sustained significant damage to the DTN (>70%). One rat was excluded from the behavioral analysis because damage to the DTN was minimal (
© 2013 American Psychological Association 0735-7044/13/$ ! 2.00 DOI: 10.1037/a0034931
The Effects of Bilateral Lesions to the Dorsal Tegmental Nucleus on Spatial Learning in Rats Jessica A. Dwyer, Matthew L. Ingram, Anna C. Snow, Christina M. Thorpe, Gerard M. Martin, and Darlene M. Skinner Memorial University of Newfoundland The head-direction (HD) signal is believed to originate in the dorsal tegmental nucleus (DTN) and lesions to this structure have been shown to disrupt HD cell firing in other areas along the HD cell circuit. To investigate the role of the DTN in spatial navigation, rats with bilateral, electrolytic (Experiment 1), or neurotoxic (Experiment 2) lesions to the DTN were compared with sham controls on two tasks that differed in difficulty and could be solved using directional heading. Rats were first trained on a direction problem in a water T maze where they learned to travel either cast or west from two locations in the experimental room. DTN-lesioned rats were impaired relative to sham controls, both early in training, on the first block of eight trials, and on the total trials taken to reach criterion. In the food-foraging task, rats were trained to leave a home cage at the periphery of a circular table, find food in the center of the table and return to the home cage. Again, DTN-lesioned rats were impaired relative to sham rats, making more errors on the return component of the foraging trip. These data extend previous cell-recording studies and behavioral tests in which rats with electrolytic DTN lesions were used, and they demonstrate the importance of the direction system to spatial leaming. Keywords: direction learning, head-direction cells, water maze, T maze, foraging task
When an animal navigates its environrnent, it relies on knowledge of the direction in which it is headed and its relative location in space (Taube, 2007). These two components of navigation are well-represented in a spatial network within the mammalian brain that contains place cells, grid cells, and head-direction (HD) cells (Fyhn, Molden, Witter, Moser, & Moser, 2004; Hafting, Fyhn, Molden, Moser, & Moser, 2005; Moser, Kropff, & Moser, 2008; Muller, 1996; O'Keefe & Dostrovsky, 1971; Taube, 1998; Taube, 2007; Taube, Muller, & Ranck, 1990a, 1990b). O'Keefe and Dostrovsky's (1971) initial report of neurons in the rat hippocampus that fired in a location-specific manner led to the suggestion that the hippocampus was crucial to the formation of spatial maps (O'Keefe & Nadel, 1978). More recent studies have revealed that place cells are part of a broader network for location that also includes grid cells of the entorhinal cortex (Fyhn et al., 2004; Hafting et al., 2005). Like hippocampal place cells, grid cells show spatially selective firing, but unlike place cells, they have multiple firing fields that form a periodic triangular array over the entire environment explored by the animal. Although place cells and grid cells provide a means of knowing our positions in the environment, to get from place to place probably requires heading information fiom HD cells (Kubie & Fenton, 2009; Moser et al., 2008; Taube,
1998, 2007). HD cells fire when an animal's head is pointed in a particular direction and, unlike place cells or grid cells, are independent of the animal's location (Taube, 2007: Taube et al., 1990a, 1990b). The preferred firing direction of HD cells can be controlled by both external (allothetic) cues (e.g., visual) and intemal (idiothetic) cues (e.g., vestibular and proprioceptive; Taube, 2007; Taube & Burton, 1995; Taube et al., 1990b). Although HD cells have been identified in various brain regions (Chen, Lin, Green, Barnes, & McNaughton, 1994; Cho & Sharp, 2001; Mizumori, Ragozzino, & Cooper, 2000; Taube, 1995; Taube et al., 1990a: Wiener. 1993; see Wiener & Taube, 2005 for review), they appear to be most abundant, and best studied, in limbic areas that form part of the classical Papez circuit (see Taube, 2007 for review). Through a combination of lesion and cell-recording studies, it has been determined that the HD signal probably originates in the reciprocal connections between the dorsal tegmental nucleus (DTN) and the lateral mammillary nuclei (LMN), and it is processed through mostly serial projections (with some reciprocal connections) from DTN^LMN^anterior dorsal nucleus of the thalamus (ADN)^postsubiculuin/parasubiculum (PoS/PaS) and retrospenial cortex (RSP; see Clark & Taube, 2012 for review). The connections between the PoS and the entorhinal cortex, the major input to the hippocampus, permit the directional signal to be integrated with location information from grid cells and place cells (see Taube, 2007 for review). Furthermore, developmental studies have revealed that HD cells exhibit adult-like firing properties earlier than place and grid cells (Langston et al., 2010; Wills, Cacucci, Burgess, & O'Keefe, 2010), leading to speculation that they may set up the entire spatial network in the mammalian brain. Although HD cells are thought to underlie an animal's sense of direction (Taube, 1998, 2007), correlations between HD cell ac-
Jessica A. Dwyer, Matthew L. Ingram, Anna C. Snow, Christina M. Thorpe. Gerard M. Martin, and Darlene M. Skinner. Department of Psychology, Memorial University of Newfoundland, St. John's, Newfoundland, Canada. This work was supported by an NSERC grant to Darlene M. Skinner. Correspondence concerning this article should be addressed to Darlene M. Skinner, Department of Psychology, Memorial University of Newfoundland, St. John's, NL AlB 3X9, Canada. E-mail: [email protected] 867
868
DWYER, INGRAM, SNOW, THORPE, MARTIN, AND SKINNER
tivity and behavior are often weak (Muir & Taube, 2002, 2004; Stackman, 2011) and, compared with the vast literature on the effects of damage to the hippocampal system on spatial navigation, relatively few lesion studies have been conducted to examine the behavioral effects of damage to the HD system. Indeed, as suggested by Stackman (2011), the lack of correlation between HDcell activity and spatial performance may be due in part to recording cell activity in brain areas that are not engaged by the behavioral task employed. He suggested an empirical approach in which a researcher first finds a behavioral task that requires an animal to use directional information, and then assess whether performance on the task depends on the activity of HD cells. Behavioral work in our lab has demonstrated a key role for direction (or orientation) to the solution of a variety of spatial problems. Through numerous experiments designed to simplify tasks so that the precise role of place, response, and direction strategies, as well as the contribution of apparatus cues versus distal cues, could be determined, we have repeatedly shown that starting orientation is important to the solution of many spatial tasks (Skinner et al., 2003; Skinner, Horne, Murphy, & Martin, 2010; Whyte, Martin, & Skinner, 2009). Using a paradigm in which the location of the apparatus changes across trials, we have shown that changes in orientation at the start points of the different maze positions are critical to the solution of place and direction problems (Peckford, McRae, Thorpe, Martin, & Skinner, 2013; Skinner et al., 2010) and facilitate response-reversal learning (Wright, Williams, Evans, Skinner, & Martin, 2009). The discrimination between starting orientations is not dependent on the view of distal visual cues from the maze, but does require some access to room cues (Peckford et al., 2013). In that study, rats were trained from two maze positions to swim in a consistent direction from choice point to goal. When the maze was later moved to two new locations, rats required to make the same response based on start-arm orientation showed no disruption in performance. The same pattem of results was obtained if the room was made dark prior to the start of each trial. However, if the rats were brought into an already darkened room then performance was disrupted. Given that HD cells are thought to underlie an animal's sense of direction, and we have developed tasks in which sense of direction or orientation is critical to the solution, we sought to determine if lesions to a component of the HD cell pathway would disrupt our behavioral measures of direction leaming. The purpose of the present study was to assess the behavioral effects of lesions to the DTN on spatial navigation. Because lesions to the DTN have been shown to disrupt the HD cell signal in other areas of the HD cell circuit, including the LMN (Bassett, Tullman, & Taube, 2007), ADN (Bassett et al., 2007), and PoS (see Taube, 2007 for review), and there is speculation that the HD signal originates subcortically in the reciprocal connection between the DTN and the LMN (Clark & Taube, 2012), we predicted that bilateral lesions to the DTN would impair performance on tasks in which directional heading is important. Very few studies have been conducted that assess the behavioral effects of DTN lesions. Although the DTN has HD cells and neurons modulated by both head direction and angular head velocity, the vast majority of the cells within the DTN fire solely in relation to the animal's angular head velocity (Taube & Bassett, 2003). Frohardt, Bassett, and Taube (2006) showed that rats with lesions to the DTN were severely impaired in a food-carrying task.
whereas rats with ADN lesions were only mildly impaired, despite the fact that the ADN has a much higher proportion of HD cells than the DTN. These results imply that some compensatory mechanism is available following damage to the ADN that is not available after DTN lesions and are consistent with the notion mentioned above that the DTN is important in generating the HD signal. A recent study by Clark et al. (2013) showed that lesions to the DTN impaired navigation based on distal landmarks using a modified water-maze procedure (see Hamilton et al, 2008). In both of these earlier studies, the rats were given electrolytic lesions to the DTN. In the present study, we assessed the effects of both electrolytic and neurotoxic lesions to the DTN.
Experiment 1 In the first experiment, rats with electrolytic lesions to the DTN were compared with sham controls on two tasks in which directional heading is important. In the first task, the two groups of rats were trained on a direction problem in a water T maze where they learned to travel either east or west from two start locations to an escape platform (Whyte et al., 2009). Previous work has shown that directional heading in the start arm, not cues at the goal location or the distal visual cues at the start location, is critical to the solution of this direction task (Peckford et al., 2013). Because other strategies such as place and response learning have been ruled out and only two maze positions are used, this task is simpler than other "pure" direction problems. For example, in the Clark et al. (2013) study, acquisition of direction learning was assessed in a task in which four different maze positions were used. Upon completion of the water T-maze task, the same rats were then trained on a food-foraging task (Whishaw & Maaswinkel, 1998), similar to that used by Frohardt et al. (2006) in which they were required to leave a home hole, find food, and then return home. Given the behavioral effects of electrolytic DTN lesions previously reported by Frohardt et al. (2006), and more recently by Clark et al. (2013), we predicted that lesions to the DTN would impair performance on both of these tasks.
Method Subjects. Twenty-four male Long-Evans rats, obtained from Charles River Company (St. Constant, Quebec, Canada) and weighing 250-300 g at the start of surgery, were used. The rats were singly housed in transparent plastic cages (45 X 25 X 21 cm) with metal lids in a temperature-controlled colony room (20 °C) and maintained on a 12-hr light/dark cycle with lights on at 7:00 a.m. All rats had continuous access to food and water in their horne cages during water T-maze training. One week prior to the foodforaging task, the rats were placed on a food-deprivation schedule to maintain their weights at 85% of their free-feeding weights. All procedures were approved by Memorial University of Newfoundland's Institutional Committee on Animal Care and followed the Canadian Council on Animal Care guidelines. Surgery. Rats were anesthetized with an intraperitoneal injection of 0.24M chloral hydrate (1 mLlWOg) and placed in a stereotaxic instrument (Model 900, Kopf, Tujunga, CA) in the skull-flat position. Additional supplements of 0.5 ml of chloral hydrate were administered as necessary. After a scalp incision was made, holes were drilled in the skull (one per hemisphere) using the bregma as
DORSAL TEGMENTAL LESIONS AND SPATIAL LEARNING
a reference point (anterior/posterior (AP); —11.7, medial/lateral (ML): ± 0.3). Sixteen rats were given lesions aimed at both the dorsal and ventral subdivisions of the DTN. A stainless steel electrode (NE-100, Rhodes Medical Instrument, Summerland, CA) was lowered 7.0 mm from the surface of the brain at a 20° angle, and 0.30 milliamps current was applied for 15 s in each hemisphere. Sham rats (n = 8) underwent the same procedure, but no current was applied. All rats were given a minimum of one week to recover before any behavioral testing. Apparatus and materials. The water T maze consisted of a Plexiglas plus maze inserted into a circular metal tank (120 cm in diameter, 31 cm high). Plexiglas walls surrounded the perimeter of the tank and both the walls and the plus maze extended 31 cm above the metal tank. The arms of the plus maze were 11.5 cm wide and 52.5 cm long. The entire apparatus was placed on a metal frame with wheels to facilitate location changes. The plus maze was converted to a T maze using a clear Plexiglas wall held in place by butterfly clips (See Figure 1). The Plexiglas wall prevented access to the arm opposite the start arm, but it did not obstruct the view of the entire tank and the surrounding room cues. The water level was kept approximately 3 cm below the top of the metal tank (28 cm deep). The water was left overnight to equilibrate with the room temperature (approximately 20 °C) and was made opaque by adding approximately 250 ml of nontoxic white Tempera paint (Rich Art Color Company, Northvale, NJ). The escape platform (11.5 cm in diameter and 26.5 cm high) was constructed from white polyvinyl chloride pipe filled with sand and attached to a Plexiglas base for stability. The escape platform was hidden from view approximately 1.5 cm below the surface of the opaque water. Water-maze training took place in a room (528 X 464 X 267 cm) located on the ground floor that had windows covering the north wall and doors located in the center of both the south and east walls. In the southwest corner of the room there was a sink. Shelves lined the west wall and half of the east wall. The southeast comer of the room contained a coat rack and a garbage can. The northwest corner had a small desk at which the experimenter sat to record data. Animal cages were arranged along a table below the shelving on the west wall. For the food-foraging task, the open field maze consisted of a wooden, white circular table (204 cm in diameter) raised 75 cm above the floor. Eight holes (11.5 cm in diameter) were evenly distributed (45° apart and centered 13.5 cm from the edge of the table) around the perimeter of the table and three white, square
West
East
Figure I. A schematic diagram of the water T-maze positions. Half of the rats were trained to go west from maze positions B and D and the other half were trained to go east from maze positions A and C. Anows indicate the correct path and solid black lines indicate a Plexiglas barrier. The entire apparatus (circular tank plus T maze) was translocated between trials.
wooden food cups ( 6 X 6 X 4 cm) were placed equidistant (50 cm) from each other in a triangular configuration at the center (See Figure 2). Large food pellets (1 g; BioServ, Frenchtown, NJ) were used to reinforce the foraging behavior. During training, a pellet was placed in one of the three food cups. The rats were held in wire-mesh cages (20 X 25 X 19 cm) for transport to the training room. The wire cages could be attached to runners beneath the holes at the periphery of the circular table. Wooden stairs were placed in the wire cage with the rat at the beginning of each trial to allow for easy access to the table. Once the wire cage was placed beneath a hole, a rat could leave the cage by climbing up the stairs and on to the table and return to the cage by climbing back down the stairs. For pretraining, the circular table was situated in the center of a ground-floor training room (680 X 507 X 267 cm) that had doors on the east and west walls and a small window on the west wall. There were cupboards, counters, and boxes along all four walls. In the northwest comer, there was a desk at which an experimenter sat to record data. The rack (150 X 50 X 165 cm), used to transport all of the rats in their wire-mesh cages, was pushed near the door along the east wall. For training on the food-foraging task, the circular table was situated in the center of a different training room (478 X 338 X 253 cm) located in the basement that had a door on the north wall. There were boxes along the west wall, a table and stool in the northwest comer of the room, and a sink in the southwest comer. In the northeast corner, there was a desk at which an experimenter sat to record data. The transport rack was pushed against the west wall. Procedure. Water T-maze training. The 24 rats (16 lesioned, 8 shams) were trained to swim in a consistent direction (half went east, half went west) to locate the hidden escape platform from two different maze positions (see Figure 1). The order of maze positions varied daily with no more than two trials in a row from the same position. Rats were brought into the training room with the water maze in three squads of eight and placed individually in transparent plastic holding cages (45 X 25 X 21 cm) that were lined with absorbent paper towel, and airanged on the table along the west wall. The remaining rats were housed in their home cages on metal racks in the hallway outside the training room. For each trial, the rats were carried in their holding cages around the water maze in a counterclockwise direction to a chair positioned at the start arm. A trial began when one experimenter (standing at the start arm for the entire trial) lowered a rat to water level in the start arm facing the Plexiglas wall of the maze. A second experimenter sitting at the desk in the northwest corner of the room recorded the arms visited by the rat and the latency (in seconds) to locate the hidden escape platform. A rat was considered to have made an arm choice when its entire body, not including its tail, was inside an arm. A correct trial was one in which the rat entered the arm containing the escape platform and climbed onto the platform, without previously visiting any other arms. After sitting on the platform for 5 s, the rat was returned to its holding cage at the start arm and carried back (in a clockwise direction) to the table along the west wall. If a rat did not locate the platform within 60 s, it was placed on the platform by the experimenter. The rats in each squad were run sequentially with an intertriai interval of approximately 10 min until each had received its eight daily trials. The rats were given eight trials per day until a criterion of 18/20 correct trials was
870
DWYER, INGRAM, SNOW, THORPE. MARTIN, AND SKINNER
Hole Food cup
Figure 2. A schematic diagram of the food-foraging apparatus. The black circles represent holes in the wooden table under which a wire-mesh cage was positioned. Only Holes 1, 3, and 5 were used as variable start points. Rats were trained to locate a large food pellet placed in one of three food cups (indicated in gray) and return to the variable home location.
reached. For those DTN-lesioned rats that did not reach this cdterion. training was stopped after 88 trials (double the number of trials taken to reach criterion by the slowest sham rat). Food-foraging task. The rats were first pretrained to leave a hole at the periphery of the table and retrieve a food pellet. To coax tbe rats from their cages, 1-g pellets were initially placed near the opening of the hole and incrementally moved further away with each exposure. Eventually, a single pellet was placed centrally on the table in a wooden cup, as in the foraging task, requiring the rats to leave the start hole and travel to the cup to obtain the food. Rats were given three such exposures per day for 2 min each until they were consistently leaving the home location to obtain food from the cup. The maximum number of trials required for any rat to become comfortable performing the task consistently was 21 (7 days). Once pretraining was completed, the apparatus was moved to a new room and all rats were trained to leave Holes 1,3, and 5 (three trials per day, one from each starting point) and find a pellet that was located in one of the three centrally positioned food cups (see Eigure 2). For each trial, the rats were caixied in their holding cages around the circular maze in a counterclockwise direction and placed below the designated start hole. A trial began when the rat emerged (all four limbs) from the wire-mesh cage, and ended when the rat returned to the cage. The wire-mesh cage was then removed from the table and the rat was returned to the metal transport rack along the west wall. The rats were run sequentially in three squads of eight with an intertdal interval of approximately 20 min. The circular table was rotated 45° every trial and washed with Sunlight soap and water to minimize scent trails. A trial was considered to be correct when the rat retrieved the pellet and immediately returned to the hole from which it started without visiting any other holes. First-choice errors were divided into three types. "Adjacenthole errors" occurred when a rat nose-poked the hole 45° to the left or right of its start hole, "memory errors" consisted of the exploration of Start Holes 1, 3, or 5 from a previous trial, and "other
errors" included visitation to any other hole. Because the rats often made more than one error per trial, we also analyzed total errors by the two groups. Each day, every rat received a tdal starting from Holes 1, 3, and 5 (in random order) going to Food Cup I, 2, or 3 (also randomized). Rats were run until they obtained a pellet from the centrally located food cups 60 times. Tdals in which the rat did not leave its cage or did not find food were not included in the analyses. Histology. Once the behavioral testing was completed, the rats were euthanized with intraperitoneal injections of 15% urethane. The brains were extracted and immediately fixed with cooled 2-methylbutane and stored at - 7 0 °C. Sections were taken at 30 (Jim and stained with cresyl violet Nissl for microscopic (14X) vedfication of the lesion sites, using the rat-brain atlas by Paxinos and Watson (1998) as a guide. To quantify the extent of tissue damage created by the electrolytic lesions, digital pictures were taken at the level of the central and pericentral subregions of the DTN (-9.3 mm postedor to bregma). We outlined and measured undamaged DTN with intact neurons and few glial cells using Image-J software (http://rsbweb.nih.gov/ij/), blind to the animal's individual performance in the behavioral testing. The total amount of tissue damage was calculated using the following equation as per Clark et al. (2013): tissue damage = average area in pixels^ of DTN in control rats — total area in pixels'^ of spared DTN tissue in lesioned rats / average area in pixels" of DTN in control rats X 100%.
Results Figure 3 diplays representative brain sections from a sham control and a DTN-lesioned rat showing both the central and pericentral subregions of the DTN, and a schematic diagram indicating the smallest and largest lesions. Based on histological analysis performed blind to the behavioral performance of the subjects, 13 of the 16 lesioned animals sustained significant damage to the DTN (>70%). Three rats were excluded from the behavioral analysis because of insufficient damage to the DTN and/or extensive damage beyond the boundaries of the DTN. Of the remaining rats, the absence of brain tissue (visible hole) and evidence of tissue damage (darker stain around lesion site and shrinkage of surrounding structures) in the area of the DTN allowed for lesion verification. Eight rats had damage that extended outside the DTN, including damage to the laterodorsal nucleus, the dorsal raphe nucleus and the central gray area. However, as there was no correlation between lesion size and trials to criterion on the water-maze task (r = —0.26, p = .40) we included these rats in the behavioral analyses. Along with the three rats excluded from all of the behavioral analyses due to damage outside the DTN, one additional lesioned rat was excluded from the food-foraging analyses, as it refused to leave the home cage dudng training. Nine of the lesioned rats that completed both behavioral tasks had complete, or near complete, damage of the DTN (>85%), and most important, had complete damage of the central subregion of the DTN. Water T maze. DTN-lesioned rats were impaired eady in training on the direction task. DTN-lesioned rats performed significantly worse than sham controls on the first block of eight trials, i(19) = 5.45, p < .05; see Eigure 4A. The difference between groups continued throughout training as lesioned rats
DORSAL TEGMENTAL LESIONS AND SPATIAL LEARNING
871
•10.04 ram
D
Figure 3. Panels A (posterior) and C (anterior) show images of a sham brain and panels B (posterior) and D (anterior) show corresponding images from a typical lesioned brain with bilateral electrolytic damage (evidenced by large hole) in the area of the DTN. The far right panel shows the anterior and posterior extent of the DTN lesions with hatch marks indicating the largest lesion and the solid shading showing the smallest lesions. The area of the DTN is outlined in white (adapted with permission from The Rat Brain in Stereotaxic Coordinates (4th ed.), by G. Paxinos and C. Watson, 1998, San Diego, California: Academic Press. Copyright 1998 by Elsevier Academic Press).
required sigtilficantly tnore trials than sham controls to reach criterion, ?(19) = 4.60, p < .05; see Figure 4B. Indeed, four of the lesioned rats failed to reach criterion on the task in twice as many trials as the slowest sham control rat. However, given that nine of the lesioned rats did eventually meet criterion, some strategy must have been available that does not depend on the integrity of the DTN. Food-foraging task. Along with the three rats excluded from all of the behavioral analyses due to damage outside the DTN, one additional lesioned rat was excluded from the food-foraging analyses, as it refused to leave the home eage during training. The data from the 60 trials in which rats found food were divided into five blocks of 12 trials for analyses. The DTN-lesioned rats showed severe impairments on the foraging task. A two-way (Group X Trial Block) ANOVA on trials correct revealed a significant effect of group, F(\, 18) = 204.7, p < .05. A two-way (Group X Trial Block) ANOVA on total errors over blocks of 12 trials also revealed a significant effect of group, F(l, 18) = 70.12, p < .05. These results confirm that lesioned rats performed much worse than sham rats (see Figure 5 A and Figure 5B). The absence of a trial-block effect suggests no improvement across trials on either of these measures. The control rats maintained a high level of performance and the DTN-lesioned rats were consistently impaired. A separate two-way (Group X
Error Type) ANOVA examining the different types of etTors made by rats in the two groups on their first hole choice revealed a significant effect of group, F(l, 18) = 204.7, p < .05 and a significant effect of error type, F(2, 36) = 13.74, p < .05. BonfetToni posttests revealed that subjects made more memory errors than adjacent {p < .05) or other (p < .01) errors, which did not differ (p > .05; see Figure 5C).
Experiment 2 The results from the first experiment add to a small but growing literature on the effects of damage to the HD system on spatial learning (e.g., Bett, Wood, & Dudchenko, 2012; Clark et al., 2013; Frohardt et al, 2006; Vann, 2005, 2011). Electrolytic lesions to the DTN produced impairments on a simple direction task in a water T maze. The advantages of this task were that only two maze positions were used, rather than four (as in Clark et al., 2013), and the confined trajectories make choices unambiguous. Although lesioned rats did show improvement on the water-maze task (i.e., reached criterion), they were severely impaired on the foraging task and did not appear to improve. These data are consistent with previous cell-recording studies and behavioral tests with DTNlesioned rats (Clark et al., 2013; Frohardt et al, 2006; see Taube, 2007 for review). In Experiment 2 we assessed whether neurotoxic
DWYER, INGRAM. SNOW. THORPE, MARTIN, AND SKINNER
872
u 4) Ü
m .2 c m 0)
DTN
Sham
Group
the same procedure but no injections were made. All rats were given a minimum of one week to recover before any behavioral testing. Apparatus and materials. The water-maze and foraging apparatus were the same as used in Experiment 1. Water-maze training was conducted in the same room as in Experiment 1. The room used for pretraining on the foraging task (523 X 439 X 267 cm) had windows on the north wall with a desk and a table and computers. There were three doors in the room, one on the west wall, one on the east wall, and one on the south wall. Shelves containing lab supplies lined the east, south and half of the west wall. Undemeath the shelving, in the southeast corner of the room were a countertop and a sink, and a table was positioned along the
12-
-©- Sham - • - DTN
108-
o s
6-
IS
o Sham
420'
DTN
Group Figure 4. (A) Mean (+ SEM) trials correct over the first eight trials for DTN-lesioned rats and sham controls on a direction task in a water T maze. (B) Mean (+ SEM) trials to criterion of DTN-lesioned rats and sham controls on the direction task.
lesions to the DTN would produce similar deficits on these two tasks. We also included a simpler, fixed-hole version of the foodforaging task. Subjects. Sixteen male Long-Evans rats, obtained from Charles River Company (St. Constant, Quebec, Canada) and weighing 250-300 g at the start of surgery, were used. The rats were housed and maintained as in Experiment 1. All procedures were approved by Memorial University of Newfoundland's Institutional Committee on Animal Care and followed the Canadian Council on Animal Care guidelines. Surgery. Rats were anesthetized with an intraperitoneal injection of 0.24 M chloral hydrate (1 mL/100 g) and placed in a stereotaxic instrument (Model 900, Kopf, Tujunga, CA) in the skull-flat position. Additional supplements of 0.5 ml of chloral hydrate were administered as necessary. After a scalp incision was made, holes were drilled in the skull (one per hemisphere) using bregma as a reference point (AP: — 11.7, ML: ± 0.3). Fight rats were given lesions aimed at the DTN and eight rats served as sham controls. Neurotoxic lesions were produced by injecting 0.3 |jil of 100 mM N-methyl-D-aspartate (Sigma Chemical, St. Louis, MO), mixed in phosphate-buffered saline, into each hemisphere using a 1 p,l Hamilton syringe and micro injection unit (Model 5000, Kopf). The syringe was lowered 7.1 mm from the surface of the brain at a 20° angle. The N-methyl-D-aspartate was injected at a constant rate over 6 min, and the needle was left in position for an additional 5 min after the injection. Sham rats (n = 8) underwent
2 3 4 Blocks Of 12 Trials
B 50-
-©- Sham - • - DTN
402 o
30-
1 ^°' 1—
102
0
3
4
Blocks Of 12 Trials
• • Memory Errors • Adjacent Errors ra Other Errors
Sham
DTN Group
Figure 5. (A) Mean ( ± SEM) trials correct over five blocks of 12 trials for DTN-lesioned rats and sham controls on the food-foraging task. (B) Mean(± 5£Af) total number of errors made over five blocks of 12 trials for DTN-lesioned rats and sham controls on the food-foraging task. (C) Mean (-f- SEM) number of first-choice memory, adjacent and other errors made over 60 trials for DTN-lesioned rats and sham controls on the foodforaging task.
DORSAL TEGMENTAL LESIONS AND SPATIAL LEARNING
west wall. The cart housing the animal cages was positioned in a space near the southwest comer of the room; the experimenter sat at a table in the northeast comer. The training room (589 X 465 X 267 cm) for the fixed-hole version of the foraging task was a large room that had been divided into two training areas with a blue curtain in the center of the room. There were two doors in this room, one on the west wall and one on the north wall. The south wall was lined with windows and the west wall was lined with shelves. There were two tables in the room; one along the west wall and one along the east wall. A long countertop (with a shelf above it) ran along the southemmost half of the east wall, and various lab supplies could be found underneath this counter. In the southwest corner of the room, there was a sink. This fixed hole task was completed in the southern half of the room and the experimenter sat by the sink in the northeast corner. The cart housing the animal cages was positioned in a space near the curtain in the center of the room. The training room for the variable-hole version of the foraging task (259 X 589 X 267 cm) had two doors, one on the north wall and one on the east wall. On the west wall was a large air conditioner. In the northwest comer of the room there was a sink and, during experimental trials, the experimenter sat by this sink. Windows lined the south wall and a computer table (with a computer) was positioned in front of these windows. The cart housing the animal cages was positioned by the door on the north wall. Procedure Water T-maze training. The rats were trained to swim in a consistent direction (half went east, half went west) to locate the hidden escape platform from two different maze positions. All details of the training procedure for this phase were identical to those outlined in Experiment 1. Training was stopped at 120 trials for those rats that did not reach criterion. Food-foraging task. The rats were first pretrained to leave a hole at the periphery of the table and retrieve a food pellet, as in Experiment 1. Because the amount of pretraining required was variable, and some rats completed pretraining days before others.
873
all rats were administered refresher trials on the day before regular training began. Upon completion of pretraining, the apparatus was moved to a new room and the rats were trained to retum to a single, fixed hole on the apparatus after foraging for a food pellet (located in one of the three food cups). During training, each rat was pseudorandomly assigned a home hole (1, 3, 5, or 7, as shown in Figure 2). The rats were given three trials/day until they reached a criterion of 9/10 trials correct. Training was stopped at 34 trials for rats in the lesioned group that did not meet criterion; this was double the number of trials required by the slowest sham rat. All other details of the training phase were identical to those of Experiment 1. Upon completion of the fixed-hole version of the task, the apparatus was moved again and the rats were trained on the variable-hole version of the task. All details of the training procedure for this task were identical to those of Experiment 1. Histology. Once the behavioral testing was completed, the rats were euthanized and the brains were processed as in Experiment 1.
Results Two rats (one sham and one lesioned rat) were euthanized due to excessive scratching around the surgical site; the sham rat had completed training in the water maze but the lesioned rat had not. A second lesioned rat died during the fixed-hole version of the foraging task. Figure 6 shows representative brain sections from a sham control and a DTN-lesioned rat. Based on histological analysis performed as per Experiment 1, six of the seven lesioned animals sustained significant damage to the DTN (>70%). One rat was excluded from the behavioral analysis because damage to the DTN was minimal (
