Behavioural Brain Research 138 (2003) 165/178. Research report

Transcription

1 Behavioural Brain Research 138 (2003) 165/178 Research report Effects of light deprivation on recovery from neglect and extinction induced by unilateral lesions of the medial agranular cortex and dorsocentral striatum T.M. Van Vleet a, S.A. Heldt a, B. Pyter a, J.V. Corwin a, *, R.L. Reep b a Department of Psychology, Northern Illinois University, De Kalb, IL 60115, USA b Department of Physiological Sciences, University of Florida, Gainesville, FL 32610, USA Received 5 April 2002; received in revised form 6 August 2002; accepted 6 August 2002 Abstract A number of previous studies have indicated that an environmental manipulation, 48 h of light deprivation (LD), produces virtually complete and permanent behavioral recovery of function from neglect induced by medial agranular cortex (AGm) lesions. LD-induced behavioral recovery from neglect is correlated with physiological changes in the dorsolateral striatum, an area that contains the projection zone of AGm efferents in the dorsocentral striatum (DCS). In this study, the behavioral effects of 48 h of LD on subjects with either unilateral DCS, AGm, or combined AGm/DCS lesions were investigated to examine whether the integrity of the DCS is crucial for behavioral recovery from neglect and whether LD will have a therapeutic effect on extinction deficits. Subjects were tested for extinction to bilateral simultaneous stimulation of the forepaws, and visual, auditory and tactile neglect. Forty-eight hours of LD failed to produce behavioral recovery from neglect in rats with DCS lesions, or a therapeutic affect on extinction deficits in any of the groups. The results of this study further support the crucial role of the DCS in recovery from neglect induced by AGm lesions and suggests that the DCS may be the crucial site for the mechanisms leading to LD-induced recovery. Further, the ineffectiveness of LD on extinction suggests that components of the neglect syndrome are dissociable and may require different therapeutic interventions. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Neglect; Light deprivation; Attention; Prefrontal; Extinction; Recovery 1. Introduction Neglect is a complex and devastating human neuropsychological disorder characterized by a failure to attend to novel or meaningful stimuli presented to the side contralateral to a brain lesion, in the absence of a primary sensory or motor dysfunction. Some manifestation of neglect is found in approximately 40% of all cases of brain damage [22]. The vast majority (80 /90%) of all cases of neglect are produced by destruction of one of the three cortical regions: the dorsolateral prefrontal cortex (area 8), the cingulate cortex, or the inferior parietal lobule [22]. The disorder goes far beyond a lack * Corresponding author. Tel.: / ; fax: / address: jcorwin@niu.edu (J.V. Corwin). of responsiveness to contralesional stimuli to include dramatic attentional and cognitive spatial deficits. The failure to respond to contralesional stimuli has devastating effects for the patients abilities to demonstrate independent daily living or to return to work, and the presence of persistent neglect leads to a poor prognosis [16,18,25]. Recovery, when it occurs, is spontaneous over the course of weeks to months, but often it is incomplete. Many patients continue to ignore or neglect contralesional stimuli for months or years post-lesion [20,23]. They are often unaware of their neurological status (anosognosia), or are unconcerned and effectively flat (anosodiaphoria) [22]. This lack of awareness and changes in motivation interfere with successful physical and occupational rehabilitation [40]. A rodent model of neglect has been developed to examine the basic mechanisms of neglect and the potential for recovery of function [2 /4,8,10]. The /02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S ( 0 2 )

2 166 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/178 symptoms of neglect in rodents following unilateral lesions of the medial agranular cortex (AGm), the rodent analog of area 8, are similar to those found in human neglect patients and include: severe neglect of visual, tactile, and auditory stimulation, extinction, allesthesia/allokinesia, and disorders of spatial processing [1,8,14,15,27,28,45]. In rats, recovery from AGm-induced neglect has been found to occur in three contexts: (1) as in humans, dopamine (DA) agonists produce acute recovery of function [10,17,19,24,26]; (2) as found in humans, rodents with neglect demonstrate spontaneous although incomplete recovery over the course of weeks to months [10,26,28]; and (3) exposure to 48 h of light deprivation (LD) at 4 h post-surgery produces dramatic immediate recovery [2,15]. Several recent studies in rodents strongly indicate that the striatum may play a critical role in neglect, and both drug-induced and spontaneous recovery from neglect produced by unilateral AGm lesions [50,53/55,58,59]. Vargo and Marshall [53,54] using immediate early gene expression (IEG) methods found that spontaneous recovery from neglect was correlated with symmetrical IEG expression in the dorsolateral striatum, and in a subsequent study [55] correlated with changes in glutamate receptors in the dorsolateral striatum. The importance of the striatum in AGm lesion-induced neglect was also suspected because apomorphine (a DA receptor agonist) can produce acute recovery [10,26], and spiroperidol (a DA receptor antagonist) can reinstate neglect [57]. The role of DA in cortical [10,12,26] and subcortical neglect [30 /34], led Fleet et al. [17] and Hurford et al. [24] to examine the effects of bromocriptine, a D2 receptor agonist, in human patients with chronic neglect. Bromocriptine produced recovery across a range of measures. When treatment was terminated, neglect worsened. Recently, Geminiani et al. [19] found that apomorphine can also produce acute recovery from neglect in humans. These findings led to a series of studies designed to investigate the role of the striatum in recovery from neglect induced by unilateral destruction of the AGm. Van Vleet et al. [50] examined the effects of unilateral destruction of the striatal projection zone of the AGm in the dorsocentral striatum (DCS) [38]. The results indicated that unilateral axon-sparing DCS lesions produced severe neglect and that the DCS operates did not demonstrate spontaneous recovery, even in subjects tested at 96 days post-lesion. Further, apomorphine did not produce a therapeutic effect in the DCS operates. In another study, Van Vleet et al. [49] examined the effects of direct infusion of apomorphine into the DCS or a laterally adjacent region of the striatum in subjects with severe neglect induced by unilateral AGm lesions. Direct infusion of apomorphine into the DCS produced a dramatic acute recovery, while infusion into the adjacent region of the striatum did not produce recovery [49]. The results of these studies suggest that the DCS may be a crucial region for recovery from neglect induced by unilateral AGm destruction. Recovery from neglect produced by unilateral AGm lesions can also be produced by an environmental manipulation, 48 h of LD. LD produces dramatic, accelerated recovery from severe neglect that lasts for the duration of behavioral testing [2,9,14,15]. Vargo et al. [59] examined the behavioral and neural correlates of LD-induced recovery and found that recovery from AGm-induced neglect was inversely correlated with behavioral activity and with the degree of asymmetry in c-fos expression in the right and left dorsolateral striata. LD decreased the lesion-induced asymmetry in the dorsolateral striatum and produced recovery. The region described as the dorsolateral striatum in the Vargo et al. [59] study encompassed much of the DCS. The Vargo et al. [59] study pointed to the potential importance of the DCS in LD-induced recovery from neglect produced by unilateral AGm lesions, but the data were correlational and included the entire dorsolateral striatum rather than the site of the AGm projections to the DCS. In this study, we directly examined the therapeutic effectiveness of LD on severe neglect produced by unilateral AGm and/or DCS lesions. In addition, we extended what is known about LD-induced recovery by examining the effects of LD on one of the major deficits associated with the neglect syndrome in humans, extinction to bilateral simultaneous stimulation. Examination of the effects of LD on extinction deficits is important because extinction deficits often persist in patients who have recovered from neglect [25]. 2. Methods 2.1. Subjects Subjects were 43 male Long-Evans hooded rats bred from stock purchased from Harlan Sprague/Dawley. Prior to the surgical procedures, the subjects were handled daily (1/2 min per day) for 21/28 days to gentle them for subsequent behavioral testing. Animals were tested between 12:00 and 17:00 h. All procedures conformed to institutional guidelines for the care and use of animals Surgical groups Prior to surgery, subjects were assigned to one of the surgical groups. There were six groups. Five groups received 48 h of LD: (1) a combined unilateral AGm and DCS lesion group (AGm/DCS; n/8); (2) a group with combined AGm and control lesions of the area laterally

3 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/ adjacent to the DCS (AGm/LSC; n/6) which acted as a control for extent of cortical and striatal damage; (3) a unilateral DCS lesion group (n/8); (4) a unilateral AGm lesion group (AGm/LD; n /5); and (5) a unilateral AGm lesion group that did not receive extinction testing (AGm; n/5). Prior studies have only examined the effects of LD on neglect [2]; therefore, this AGm group, which was taken from a prior study from our laboratory [2], was used to control for the possible effects of combined neglect and extinction testing on the therapeutic effectiveness of LD. A sixth group which received a unilateral AGm lesion was tested for both neglect and extinction, but did not experience LD (AGm/noLD; n /8). This group was included in the analysis of the extinction data only, and served as a baseline comparison group to examine the affects of LD on extinction. Finally, to examine the possibility that the effective dose of LD may differ for the AGm and the DCS, three additional unilateral DCS operates were given 72 h of LD. A prior study has indicated that 72 h rather than 48 h of LD is necessary to consistently produce recovery in 24-month-old AGm operates with severe neglect [7] Preoperative extinction testing In order to assess for a preoperative paw preference, subjects were given a test commonly used to examine extinction to bilateral simultaneous stimulation [29,43,51]. The subjects, while still in their home cage, were placed on a table in a room illuminated by dim red light for 1 min of adaptation. After adaptation, round adhesive-backed papers (1.3 cm diameter labels) (Avery International) were attached bilaterally to the radial aspect of the wrist. The distal part of the label covered approximately 1 mm of the hairless part of the forepaw. After both labels were attached, they were pressed simultaneously, and the animal was replaced in its home cage. Latency to contact and latency to remove each label were recorded. If the animal failed to remove one or both of the labels within 120 s of attachment, the trial was terminated. Each animal was given five trials per test session, with each trial separated by 1 min Surgical procedures AGm lesions A detailed description of the surgical procedures can be found in Burcham et al. [2,4]. Briefly, animals were anesthetized using sodium pentobarbital (50 mg/kg, i.p.). Under aseptic conditions, a midline incision was made and the skull revealed. A skull window extending from 5 mm rostral to /2 mm caudal to bregma and 2 mm lateral to the sagittal sinus was removed on the intended side of the lesion. The AGm was then removed via gentle subpial aspiration. Once hemostasis was achieved, Gelfoam was placed in the wound and the incision closed with sterile wound clips and treated with Neomycin ointment. The subjects were kept warm and monitored prior to behavioral testing DCS and LSC lesions Animals were anesthetized using sodium pentobarbital (50 mg/kg, i.p.). When unresponsive (absence of corneal reflex and responsiveness to a mild tail pinch), the subjects were placed in a stereotaxic apparatus using blunt-tipped ear bars. Under aseptic conditions, a small hole was drilled in the skull at 0.5 mm rostral and 3.4 mm lateral to bregma. Drugs were infused using a programmable infusion pump (Stoelting) attached to a 0.5 ml, 26-gauge Hamilton syringe. For the DCS group, the cannula was lowered 4.2 mm ventral to the dura. For the LSC group, the hole was drilled at 0.5 mm rostral, 4.4 mm lateral relative to bregma, and the cannula lowered to 5.2 mm ventral to the dura. The 0.1 ml dose of NMDA was infused at a rate of 0.1 ml/2 min, and allowed to diffuse for an additional 4 min before the syringe was removed. The incision was closed with sterile wound clips, treated with neomycin ointment, and the subjects kept warm and monitored prior to being returned to the colony Combined lesions In the AGm/DCS and AGm/LSC groups in which the subjects received both surgeries in the same session, the subjects received the appropriate striatal lesion first, followed by subpial aspiration of the AGm as described earlier for each of the individual procedures. The subjects tolerated the combined procedures quite well, and could be tested for neglect at 4 h post-injection Pre-LD behavioral testing Circling behavior The subject was placed in its home cage on the testing platform, and the number of ipsi- and contralesional turns was counted to the nearest 1/2 turn for a 2 min period prior to behavioral testing for neglect [10,15]. A 1/2 turn was recorded when both of the subject s hind feet crossed over an imaginary line 908 from the starting position Orientation testing All testing were conducted during the light phase of the light/dark cycle in a room with standard overhead fluorescent lighting. Orientation testing was a modified version of that developed by Crowne et al. [15] and was designed to reflect simple bedside testing for neglect in humans. These procedures have been used in a number of studies from my laboratory [2,4,50] and others [59]. After the 2 min assessment of circling behavior, the animal was taken out of its cage and placed directly on

4 168 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/178 the test platform marked to delineate 08, 308, 458, and 608 angles in either direction from a central line running the length of the testing board. The subject was gently restrained by hand from behind without restricting head movement, and aligned with the centerline. Stimuli were presented only when there was no evidence of struggling, no asymmetry of body posture, and when the head was oriented in direct line with the body. Typically, the animal s body had to be realigned several times during testing. The early extensive handling minimized struggling and freezing. Visual, tactile, and auditory stimuli were presented in turn. The visual stimulus consisted of the presentation of a silver metallic rod 10.0 cm in length (8 mm in width), which was waved in a small circle (approximately 5.0 cm in diameter) five times within the animal s visual field at a distance of 7.5/10.0 cm from the animal. Care was taken not to contact the vibrissa while waving the metallic rod. The auditory stimulus was a single 114 db (SPL) click generated by a clicking device held at mid-body, approximately 5 cm from the subject. The tactile stimulus was a single caudal-to-rostral stroke through the vibrissa with a 15 cm Puritan applicator (Harkwood Products Co., No. 807). Although such stimulation has a visual component, testing under red light conditions yielded virtually identical results (Corwin, unpublished). Three cycles of testing comprised one test session. One cycle consisted of a single presentation of each of the three stimuli to each body side in turn. Stimuli were presented in the order: visual, tactile, and then auditory. We have found that order of presentation does not influence performance (Corwin, unpublished). The experimenter rated the degree of head turning toward or away from (allesthetic responding) the stimuli as measured by the position of the tip of the snout over the test platform markings. A head turn of less than 308 received a zero score, between 308 and 458 as 1.0, between 458 and 608 as 1.5, and greater than 608 as 2.0. Orientations later than 2 s after stimulus presentation received a zero score. Orientations to the visual stimulus after the third revolution (3 s) received a maximum score of 1.5. The maximum score for each body side is 6.0 for each of the three modalities, 18.0 in total. In prior studies, this rating scale produced an interrator reliability of 1.0 for the direction of orientation, and above 0.9 for the magnitude of orientation [10]. The experimenter testing the animals was blind with respect to the group affiliation of the subjects. A total neglect ratio was derived from the formula: total contralesional responsiveness/total ipsilesional responsiveness, and collapsed across all modalities [2,4,10]. The neglect ratio is a measure of relative responding to each body side and takes into account total responsiveness; it indicates asymmetries in orientation behavior. Neglect ratios were also calculated to compare the non-neglected to the neglected body side for the visual, tactile, and auditory modalities. Because neglect ratios give no indication of whether an asymmetry in orientation results from lower contralesional (non-neglect side), or higher ipsilesional (neglect side) scores separate analyses of raw scores for ipsi- and contralesional responding was conducted to assess prevs. post-manipulation responsiveness [57] Allesthesia/allokinesia Responses to the inappropriate side, away from the side of stimulation (allesthetic responses), were rated identically. All behavioral testing occurred between 07:00 and 19:00 h of the light portion of the 12/12 day/night cycle. All LD subjects were tested for neglect 4 h after the appropriate surgical procedures. As in prior studies, we used animals with severe neglect: (total contralesional responsiveness/total ipsilesional responsiveness) 5/0.33, in order to conservatively test the effects of LD [2,4,9]. The time post-injection when neglect is first assessed (4 h) was chosen with two concerns in mind: (1) that as short an interval post-surgery as possible be chosen that would allow for behavioral assessment of neglect and (2) severe neglect demonstrated at the first test would predict severe neglect on subsequent tests. Both these criteria have been met in previous studies of LD [2,4,9]. Severe neglect at 4 h post-injection has invariably predicted severe neglect at 48 h post-surgery [2,9]. In the AGm and AGm/LSC groups, our criteria for severe neglect was met by approximately 80% of all operates. In contrast to prior studies [2,9], we have found that DCS and AGm/DCS subjects consistently exhibit severe chronic neglect at 48 h post-surgery regardless of the severity of neglect during the Pre-LD test at 4 h postinjection [51]. This may due to the possibility that the neurotoxicity produced by NMDA requires a longer time course to produce severe neglect. In this study, only 50% of the subjects in the AGm/DCS and DCS groups demonstrated severe neglect at 4 h post-injection Light deprivation The environmental manipulation took place in a 3 m /3 m windowless room. At any one time, one to three animals were in the experimental room. Care was taken to ensure that sufficient food and water was given for the duration of the manipulation. The door was sealed and all extraneous light was blocked. It was typical that different groups experienced the environmental manipulation together to prevent order effects Post-LD behavioral testing Orientation testing Following the environmental manipulation, the door to the room was opened, the lights turned on, and the

5 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/ subjects allowed at least 15 min of adaptation to the lighted room. After adaptation, the subjects were tested for neglect in a fashion identical to that described for the orientation test at 4 h post-injection. Thereafter, all subjects were kept in the colony under a 12:12 light/dark cycle. The experimenter was blind with respect to the subjects group affiliation. Subjects were tested twice per week and the effects of LD on neglect were evaluated for 2 weeks Extinction testing Extinction testing procedures were identical to those given prior to surgery. Extinction was considered to be a tendency to touch and/or remove the tab attached to the ipsilesional wrist before touching and/or removing the contralesionally applied tab during an individual trial. Therefore, analyses of data on the extinction task took place on the following measures: (1) frequency of touching the ipsilesional tab first, (2) frequency of removing the ipsilesional tab first, (3) frequency of touching the contralesional tab first, and (4) frequency of removing the contralesional tab first. Extinction was assessed on alternate days when the subjects were not being tested for orientation (neglect) [29]. The experimenter was blind with respect to the group affiliation of the subjects Histological procedures After behavioral testing was completed, the subjects were given an overdose of sodium pentobarbital (65 mg) and when totally unresponsive (absence of a corneal reflex, unresponsiveness to tail pinch, and cessation of respiration), intracardially perfused with normal saline, followed by 10% formalin. The brain was removed from the skull, placed in 10% formalin for at least 3 days, and then into 30% sucrose /formalin until the brain sank. The brain was frozen and sectioned in the coronal plane at 50 mm. For the AGm operates every tenth section through the extent of the lesion and every fifth section through the thalamus was saved, mounted, and stained with cresyl-violet. All lesions were examined to determine the extent of damage to adjacent areas and the AGm. The thalamus was examined for any signs of calcification, gliosis, or shrinkage of nuclei. For the DCS and LSC operates every section through the extent of the lesion was saved, mounted, and stained with cresyl-violet. Lesion sizes were measured by tracing the extent of lesion and areas of gliosis through an image analysis program (Optimus, BioScan). Lesion extents were traced onto standard brain diagrams [37]. All brains were identified by an arbitrary number, and not associated with a particular group. All histological processing and analyses were performed prior to the behavioral analyses in order to assure blindness. 3. Results 3.1. Histology The histological analysis revealed that the unilateral AGm lesions were virtually identical to those in other studies of neglect [2,4,10,14,26,56,57]. As indicated in Fig. 1, in addition to the removal of AGm, typically there was also some damage to the lateral agranular cortex, the dorsal cingulate cortex, and the cingulum bundle. The size of the unilateral AGm lesions did not differ among the groups and were virtually identical to those in prior studies [2,4,9]. Likewise, DCS lesion size was virtually identical across the groups and was virtually identical to those in a prior study of the effects of DCS lesions [50]. The DCS lesions in the AGm/DCS and the DCS groups destroyed a significant portion of the projection zone of the Agm; however, there was very minor damage produced by the injection procedure itself. The LSC lesions in the AGm/LSC group were virtually identical in size to the DCS lesions. Image analysis revealed no significant difference in lesion extent between the AGm/DCS group and the AGm/ LSC group; therefore, any behavioral differences cannot be explained by differences in lesion size Neglect A total neglect ratio was derived from the formula: contralateral total neglect/ipsilateral total neglect [2,4,10,50]. The total neglect ratio is a measure of relative responding to each body side and takes into account total responsiveness. The data from the five tests, which included a Pre-LD test (Test 1) and 2 weeks of post-operative testing (Tests 2/5), were analyzed. A Group /Tests ANOVA comparing the total neglect ratios for the AGm [2], DCS, AGm/LSC, AGm/DCS, and the AGm/LD groups revealed no significant main effects (all P /0.05), but did indicate a significant Test/Group interaction (F(16,104) /2.138, P B/ 0.012). In order to explore the significant interaction, individual one-way ANOVAs were done to compare the groups at each of the five tests. The results, as illustrated in Fig. 2, indicated significant differences among the groups for Tests 2/5 (all P B/0.001). The lack of a significant difference on Test 1 (Pretest) reflects the screening for severe neglect found in all but the DCS groups. Individual comparisons using LSD analyses revealed that on Test 2, both the DCS and AGm/DCS groups, despite the absence of Pre-LD neglect, demonstrated significantly more neglect than both the AGm and AGm/LD (all P B/0.01), but did not differ from each other or the AGm/LSC group (all P /0.05). On Tests 3 /5, DCS and AGm/DCS groups demonstrated significant neglect when compared with the AGm/LD,

6 170 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/178 Fig. 1. Maximum (dark shading) and minimum (light shading) extents of the lesion for individual subjects in the AGm, DCS, AGm/DCS, AGm/ LSC, AGm/LD, and AGm/noLD groups. AGm, and the AGm/LSC operates (all P B/0.01), but did not differ from one another. These results indicate that LD had an immediate and lasting therapeutic effect in the AGm and AGm/LD groups, which was observed immediately following LD. In the AGm/LSC group, the therapeutic effects of LD did not emerge until approximately 48 h after LD treatment. In contrast, LD did not induce recovery for neglect produced by unilateral lesions of the DCS or combined unilateral lesions of the AGm/DCS. The fact that the AGm/LSC group demonstrated LD-induced recovery further suggests that damage to the DCS, and not striatal damage in general, prevented LD-induced recovery in the DCS and AGm/DCS groups. The DCS subjects that received 72 h of LD did not demonstrate any evidence of behavioral recovery from severe neglect. The absence of LD-induced recovery in the DCS and AGm/DCS groups was not due merely to insufficient duration of exposure to LD because neither 48 h nor 72 h of LD was effective.

7 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/ Fig. 2. Mean total neglect ratios in the AGm, DCS, AGm/DCS, AGm/LSC, and AGm/LD groups as a function of test. A ratio of 1.0 represents symmetrical responding. (') DCS and AGm/DCS groups were significantly different from both the AGm and AGm/LD groups (P B/0.05). (w) DCS and AGm/DCS groups were significantly different from all other groups (P B/0.01). Error bars represent standard errors. Because LD-induced changes in neglect ratios give no indication of whether an asymmetry in orientation is due to an increase in contralesional (neglect side) responding, a decrease in ipsilesional (non-neglect side) responding, or both, separate analyses of ipsi- and contralesional raw scores were conducted. Group / Test ANOVAs were done to evaluate possible group differences in ipsi- and contralesional responsiveness across tests. For ipsilesional responding, the results indicated a significant group effect (F(4,30) /2.862, P B/0.04). Post hoc analysis indicated that, collapsed across test, the AGm/DCS and AGm/LSC groups demonstrated significantly greater ipsilesional responding than the other groups (all P B/0.03). For contralesional responding, the results indicate a significant group effect (F(4,30) /16.665, P B/0.001) and Group /Test interaction (F(4,30) /19.250, P B/ 0.001). To further evaluate the Group/Test interaction for contralesional responding, one-way ANOVAs were performed to compare contralesional responding among groups at each of the five tests. The results, as illustrated in Fig. 3, indicated significant differences among the groups for Tests 1/5 (all P B/0.004). Individual comparisons using LSD analyses revealed that for Test 1 (Pre-LD), the DCS group exhibited more contralesional responding than all other groups (all P B/0.001). As indicated earlier in Section 2, immediate post-operative neglect was not exhibited in 50% of the DCS operates. For Test 2, LSD analysis revealed the following: the AGm/DCS and the DCS groups demonstrated significantly less contralesional responding than the AGm and AGm/LD groups (all P B/0.001), the AGm/DCS group demonstrated significantly less contralesional responding than the DCS and AGm/LSC groups (all P B/0.05), and the AGm/LSC group exhibited reliably less contralesional responding than the AGm and AGm/LD groups (all P B/0.02). Finally, on Tests 3/5, both the DCS and AGm/DCS groups demonstrated significantly less contralesional responding than the AGm, AGm/ LSC, and AGm/LD operates (all P B/0.02). Taken together, the results of the raw scores parallel the results of the neglect ratio and indicate that LD produced dramatic and lasting recovery from severe neglect induced by unilateral AGm lesions, although delayed in the AGm/LSC group. LD did not have an effect on neglect produced by unilateral lesions of the DCS or combined unilateral lesions of the AGm/DCS. Furthermore, the raw score analyses indicated that recovery produced by LD was primarily the product of an increase in contralesional (previously neglected side) responding, as found in prior studies [2,9]. These results are also consistent with the findings of Vargo et al. [59], which indicated that the symmetry of striatal functional activity, rather than a specific change in the lesioned hemisphere, was correlated with recovery from neglect Modalities In order to evaluate the effects of LD on the individual modalities, individual Group/Test ANOVA were done using neglect ratios for the visual, tactile, and auditory modalities. The results are presented graphically in Fig Visual The analyses of the visual neglect ratios revealed a significant main effect for group (F(4,29) /10.556, P B/ ) and a significant Group /Test interaction

8 172 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/178 Fig. 3. Mean contralesional raw scores for the AGm, DCS, AGm/DCS, AGm/LSC, and AGm/LD groups as a function of test. A score of 18.0 represents complete responsiveness across all modalities. (') DCS and AGm/DCS groups were significantly different from both the AGm and AGm/ LD groups (P B/0.001). (w) DCS and AGm/DCS groups were significantly different from all other groups (P B/0.02). (#) DCS is significantly different from the AGm, AGm/LD, and AGm/LSC groups (P B/0.001). Error bars represent standard errors. (F(16,29) /2.721, P B/0.001). Subsequent one-way AN- OVAs comparing groups at each level of test revealed significant differences at Tests 2/5 (all P B/0.02). Post hoc LSD analyses comparing the groups revealed that, on Test 2, the AGm/DCS group demonstrated more severe neglect than the DCS, AGm, and AGm/LD groups (all P B/0.02), and that the DCS and AGm/ LSC groups demonstrated more severe neglect than the AGm/LD group (all P B/0.03). For Test 3, the AGm/ DCS and DCS groups exhibited significantly greater visual neglect than the AGm, AGm/LSC, and AGm/LD groups (all P B/0.001). The AGm/DCS and the DCS groups exhibited significantly greater visual neglect than both the AGm and AGm/LSC groups on Test 4 (all P B/0.001), but not the AGm/LD group (P /0.05). On Test 5, the DCS group exhibited significantly greater visual neglect than all other groups (all P B/0.01) except the AGm/DCS group (P /0.05). In general, the results of the visual orientation tests indicate that, relative to AGm/DCS and the DCS groups, LD produced recovery from visual neglect in the AGm, AGm/LSC, and AGm/ LD groups Tactile The Group /Test analysis of the ratios for tactile neglect revealed a significant main effect for group (F(4,17) /17.371, P B/0.001) and a significant Group /Test interaction (F(16,29) /3.649, P B/ ). Individual one-way ANOVA comparisons among the groups at each level of test revealed significant group differences at Tests 2 /5 (all P B/ 0.001). Post hoc LSD analyses comparing the groups at each test revealed that on Test 2, the AGm/DCS and DCS groups exhibited significantly greater tactile neglect than both the AGm and AGm/LD groups (all P B/ 0.001), but did not differ from each other or the AGm/ LSC group (P /0.05). On Tests 3/5, the AGm/DCS and DCS groups exhibited significantly greater tactile neglect than the AGm, AGm/LSC, and AGm/LD groups (all P B/0.02), but did not differ from one another. Overall, these results indicate that relative to AGm/DCS and the DCS groups, LD produced recovery from tactile neglect in the AGm, AGm/LSC, and AGm/ LD groups Auditory The Group /Test analysis for the auditory modality revealed a significant effect for group (F(4,14) /8.366, P B/0.001) and a significant Group /Test interaction (F(16,29) /2.565, P B/0.005). One-way ANOVA analyses were conducted to examine the differences among the groups at each of the tests which indicated significant group differences on Tests 2 /5 (all P B/0.03). Post hoc LSD analyses revealed that the AGm/DCS and DCS groups exhibited significantly greater auditory neglect than the AGm group (P B/0.03), but did not differ from each other or the AGm/LD or AGm/LSC groups (all P /0.05) on Tests 2 and 3. On Test 4, the AGm/DCS and DCS groups exhibited significantly greater auditory neglect than the AGm and AGm/LD groups (P B/0.01), but did not differ from each other or the AGm/LSC group (P /0.05). On Test 5, the AGm/ DCS and DCS groups exhibited significantly greater auditory neglect than both the AGm and AGm/LSC

9 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/ Fig. 4. Individual modality neglect ratios in the AGm, DCS, AGm/DCS, AGm/LSC, and AGm/LD groups as a function of test. A ratio of 0.0 represents symmetrical responding. (w) DCS and AGm/DCS groups were significantly different from all other groups (P B/0.02). (') DCS and AGm/DCS groups were significantly different from both the AGm and AGm/LD groups (P B/0.01). Error bars represent standard errors. (#) DCS is significantly different from the AGm, AGm/LD, and AGm/LSC groups (P B/0.05). (*) DCS and AGm/DCS groups were significantly different from the AGm group (P B/0.05). (^) AGm/DCS group is significantly different from the DCS, AGm, and AGm/LD groups (P B/0.02), and that the AGm/LD group is significantly different from the DCS and AGm/LSC groups (P B/0.03). (/) DCS and AGm/DCS groups were significantly different from the AGm and AGm/LSC groups (P B/0.02). Error bars represent standard errors. groups (all P B/0.02), but did not differ from each other or the AGm/LD group (all P /0.05). Overall, LD was effective in eliminating severe auditory neglect in the AGm group, and to a lesser extent the AGm/LSC and AGm/DCS groups. In general, the results obtained in the individual modalities corroborate the findings from the overall neglect ratio analyses. They also support findings from prior studies of neglect produced by unilateral AGm destruction [10,26,56] and unilateral lesions of the DCS [50,52]. In this study, 48 h of LD was particularly effective in producing recovery from severe neglect in the visual and tactile modalities. The failure to find significant auditory neglect in the AGm/LSC and AGm/ LD groups is not uncommon. Unilateral AGm or DCS operates may demonstrate some variability of neglect symptoms within the individual modalities [9,50,52] Extinction Contacts In order to examine the potential differences in frequency of initial contact, a Group /Test /Response (ipsi- and contralesional) ANOVA was done. The results revealed a main effect for response (F(1,29) /18.002, P B/0.001) and an insignificant effect for group (F(4,29) /1.783, P /0.05). Significant Group /Response (F(4,29) /5.341, P B/0.002) and Test/Response (F(6,29) /17.421, P B/0.001) interactions were also obtained. In order to examine the Group /Response interaction, individual analyses were done to examine group differences for each side of response (ipsi- and contralesional). For ipsilesional contacts, there was no significant effect of group (F(4,29) /2.585, P /0.05). In

10 174 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/178 Fig. 5. (A) Mean frequency of contralesional contacts in the AGm, DCS, AGm/DCS, AGm/LSC, AGm/LD, and AGm/noLD groups as a function of test. The maximum number of potential contacts is 5.0. With the exception of the DCS group, all groups demonstrated a significant and lasting reduction in the frequency of contralesional contacts. Error bars represent standard errors. (B) Mean frequency of contralesional removals in the AGm, DCS, AGm/DCS, AGm/LSC, AGm/LD, and AGm/noLD groups as a function of test. The maximum number of potential removals is 5.0. With the exception of the DCS and AGm/DCS groups, all groups demonstrated a significant and lasting reduction in the frequency of contralesional removals. Error bars represent standard errors. contrast, a significant group effect was obtained for contralesional contacts (F(4,29) /7.162, P B/0.001). Pairwise comparisons revealed that the frequency of contacting the contralesional tab first was significantly greater in the DCS operates when compared with all other groups (all P B/0.007), reflecting the absence of an extinction deficit. In addition, it was also found that the AGm/LSC group exhibited significantly fewer contralesional contacts than the AGm/noLD group (P B/0.04). No other differences were detected among the AGm/ nold, AGm/DCS, AGm/LSC, and AGm/LD groups (all P /0.05). To evaluate the Test/Response interaction, separate comparisons were done for ipsi- and contralesional contacts collapsed across group. The results indicated that both ipsilesional (F(6,24) /6.378, P B/0.017) and contralesional (F(6,24) /16.701, P B/0.0001) initial contacts changed significantly over tests. Pairwise comparisons revealed the frequency of contacting the contralesional tab was greater for Test 1 (Pretest) than Tests 2 /7 (all P B/0.0001). For ipsilesional contacts, the frequency on Test 1 (Pretest) was significantly smaller than Tests 2/7 (all P B/0.0001). Analyses of the Test/Response interaction indicated that when groups are analyzed together, all operates displayed a significant increase in ipsilesional contacts and significant decrease in contralesional contacts after surgery. However, the Group /Response interaction revealed that the DCS operates, unlike the other groups, did not change their preoperative contralesional paw preference and thus displayed significantly more contacts to the contralesional side than the other groups. Taken together, the results suggest that all groups, except the DCS, exhibited a post-surgical change in contact paw preference, indicative of an extinction deficit.

11 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/ In order to determine whether individual groups displayed a change in paw preference for contact, a series of within-subject analyses were done which separately examined contra- and ipsilesional frequency of contact for each group across tests. With regard to ipsilesional tab contacts, the AGm/noLD, AGm/LD, AGm/DCS, and AGm/LSC groups demonstrated a significant test effect (all P B/0.02). In these groups, paired sample t-tests were done to compare Test 1 (Pretest) to subsequent Post-LD tests. For the AGm/ nold group, the results indicated that Test 1 differed significantly from Tests 2, 3, 4, and 6 (all P B/0.03); an indication that ipsilesional tab contact increased following the AGm lesion. In a similar fashion, the AGm/DCS group significantly increased ipsilesional contacts on Tests 2 and 3, as did the AGm/LSC group on Test 3 (all P B/0.01). For the DCS group, the analysis revealed no significant change in ipsilesional contacts. The analysis for the contralesional contacts revealed a significant test effect for the AGm/noLD, AGm/LD, AGm/DCS, and AGm/LSC groups (all P B/0.008). As seen in Fig. 5A, the analyses comparing Test 1 with the subsequent tests showed that the AGm/noLD, AGm/ LD, AGm/DCS, and AGm/LSC lesions produced a significant decrease in contacts to the contralesional preoperatively preferred paw on Tests 2 /7 (all P B/ 0.04). Taken together, the within-subject analyses indicated that all groups except the DCS group demonstrated extinction post-surgically, as evidenced by a significant increase in ipsilesional contacts and a significant decrease in contralesional contacts. Further, results from this study support recent findings from our lab which demonstrate that DCS lesions alone do not produce extinction [52] Removals A Group/Test/Response (ipsi- and contralesional) ANOVA was done to examine the pattern of tab removals. The results indicated that there were no significant main effects (all P /0.23), but there was a significant Test /Response (F(6,150) /8.372, P B/ ) interaction. The Response/Group interaction, although not significant, was noteworthy (F(4,25) / 2.686, P B/0.054). In order to evaluate the Test/Response (ipsi- and contralesional) interaction, individual comparisons were done for ipsi- and contralesional removals collapsed across group. The results indicated that both ipsilesional (F(6,24) /10.344, P B/0.0001) and contralesional (F(6,24) /8.996, P B/0.0001) removals changed significantly over the duration of testing. Follow-up comparisons revealed that when, collapsed across all groups, Test 1 differed significantly from Tests 2/7 for both ipsilesional (all P B/0.005) and contralesional (all P B/ 0.008) removals. These results suggest that, when collapsed across all groups, there was a significant increase in ipsilesional removal and a significant decrease in contralesional removal after surgery. In order to compare more directly ipsi- and contralesional removals, additional within-subject analyses were done to separately examine contra- and ipsilesional responses for each group across the tests. With regard to ipsilesional tab removals, only the AGm/DCS and AGm/LD groups demonstrated a significant test effect (all P B/0.049). Paired sample t-tests comparing Test 1 (Pretest) to subsequent post-surgical tests indicated that the number of Pretest ipsilesional tab removals in both the AGm/LD and AGm/DCS groups significantly increased on post-surgical Tests 2, 3, 4, 6, and 7 (all P B/0.04). With regard to the contralesional tab removals, the AGm/noLD, AGm/LD, and AGm/DCS groups demonstrated a significant test effect (P B/0.002). Paired sample t-tests comparing Test 1 (Pretest) to subsequent post-surgical tests indicated that the number of contralesional tab removals in the AGm/noLD, AGm/LD, and AGm/DCS lesion groups decreased post-surgically (all P B/0.04). Overall, the removal within-subject analyses indicated that the AGm/DCS, AGm/LD, and AGm/noLD groups exhibited significant extinction, as evidenced by a significant change in paw preference post-surgery. The AGm/DCS and AGm/LD groups demonstrated a significant increase in ipsilesional removals and a significant decrease in contralesional removals post-surgically. The AGm/noLD groups also demonstrated a significant decrease in contralesional responding post-surgery (see Fig. 5B). The results also corroborate the findings from the contact analyses. The failure of the AGm/LSC group to demonstrate any significant change in ipsi- or contralesional removals is likely due to a somewhat lower pretest removal frequency in this group, as illustrated in Fig. 5B Allesthesia/allokinesia Responses to the inappropriate side, away from the side of stimulation (allesthetic/allokinetic responses), were rated identically to appropriate orientations. Because of the large number of zero scores, the total number of allesthetic responses for the ipsi- and contralesional sides were compared across the initial five tests for each of the five groups using a Wilcoxon Signed-Ranks test. There were no significant differences between ipsi- and contralesional allesthesia/allokinesia responding in any of the groups (all P /0.05) Circling Wilcoxon Signed-Ranks tests were conducted to compare the amount of ipsilesional vs. contralesional circling in the five groups. The analyses of circling behavior revealed no significant differences between

12 176 T.M. Van Vleet et al. / Behavioural Brain Research 138 (2003) 165/178 ipsi- and contralesional circling within any of the groups (all P /0.05). Thus, any neglect or extinction deficits cannot be explained by a circling bias or postural asymmetries [44]. 4. Discussion Crowne et al. [14,15] found that animals that experienced eye closure or 48 h of LD did not demonstrate neglect following unilateral AGm lesions. In both studies, LD was effective in producing virtually complete sparing from AGm-induced neglect. Further, they were able to demonstrate that the sparing of function was due to the therapeutic effects of LD and not the result of auditory or activity changes in the colony [15]. Similarly, LD has also been found to produce sparing and recovery of function from sensory motor deficits produced by lesions of the lateral hypothalamus [21,42]. However, LD was not effective in producing sparing of neglect induced by unilateral lesions of the superior colliculus [15]. Corwin and Vargo [9] extended the findings of Crowne et al. on sparing of function [14,15] by demonstrating that LD can produce dramatic recovery from severe neglect if administered within 4 h. Most recently, Vargo et al. [59] have suggested that the crucial site for LD-induced recovery may be the dorsolateral striatum, by demonstrating that recovery was correlated with alterations in functional activity in this region. The results of this study extend these findings and indicate that the striatal projection zone of the AGm, the DCS, may be a crucial site within the dorsolateral striatum for the mechanisms of LD-induced recovery from neglect induced by unilateral AGm destruction. Combined unilateral destruction of the AGm/DCS prohibited recovery, whereas combined unilateral AGm/LSC lesions equivalent in size, but outside of the projection zone of AGm did not prohibit the therapeutic effects of LD on severe neglect. These findings support prior studies which indicate that the DCS may play a critical role in recovery from neglect induced by unilateral AGm lesions. In studies examining spontaneous recovery, AGm-induced neglect and recovery were correlated with changes in immediate gene expression in the dorsolateral striatum [54]. In a subsequent study, Vargo and Marshall [55] found that changes in NMDA and kainate receptors in this same region were correlated with spontaneous recovery from AGm-induced neglect. Recent observations also support the role of the DCS in acute drug-induced recovery from neglect. Systemic administration of apomorphine has been found to produce acute recovery from AGm lesioninduced neglect [10,26]. The likely site of action for the therapeutic effects of apomorphine is the DCS [49]. Direct infusion of apomorphine into the DCS in rats with severe neglect produced by AGm lesions produced virtually the same therapeutic effects as systemic injections [49]. Infusion into the more lateral striatum did not produce a therapeutic effect [49]. Further, it has been demonstrated that apomorphine is ineffective in producing recovery from severe neglect in unilateral DCS operates [50]. These findings on the role of the DCS in spontaneous and drug-induced recovery, when taken together with the present results, indicate that the integrity of the DCS may be necessary for recovery from AGm-induced neglect. Recent anatomical findings in rats indicate that the DCS is a convergence zone for projections from cortical association areas including the AGm, PPC, and the ventrolateral orbital cortex [5,39]. Damage to any one of these interconnected cortical regions produces neglect and deficits in spatial processing [2,4,8,11,14,15,26 /29]. In addition, the DCS has been implicated as a multimodal convergence region of striatum [35]. Previous work in primates has also suggested that interconnected cortical areas, that have been implicated in spatial and attention behavior, are likely to have converging corticostriatal projections [36,46]. The importance of corticostriatal relationships in rats has also been implicated in attentional deficits following bilateral lesions of striatum [41], or via disconnection studies [6]. A second major finding of this study is that the therapeutic effects of LD do not extend to extinction. In the same subjects that demonstrated recovery from neglect as a result of exposure to LD, there was no evidence that the effects generalized to extinction. For example, the AGm and AGm/LSC groups demonstrated dramatic LD-induced multimodal recovery from severe neglect by the second test after LD, but a severe extinction deficit was unaffected. These findings strongly suggest that deficits associated with the neglect syndrome likely have distinct or dissociable anatomical and pharmacological substrates [47,48,52], and that the eventual treatment of these components of the neglect syndrome will be quite complex [25]. Recent pilot data from our laboratory supports this contention by demonstrating that apomorphine produces a therapeutic effect on neglect but not extinction in rats with unilateral AGm lesions (Pyter et al., unpublished). These findings provide support for the dissociation of neglect symptoms as found in an earlier study, which compared the effects of rostral and caudal AGm lesions. Unilateral destruction of the rostral AGm produced significant allesthesia/allokinesia without severe neglect, whereas caudal AGm lesions produced severe neglect without significant allesthesia [26]. Taken together, these studies suggest that the rat model may be useful for determining the neural substrates, which underlie the specific deficits that comprise the neglect syndrome. The present findings have potential relevance for the treatment of neglect. Behavioral treatments rarely gen-