## Abstract

Recently, we demonstrated that neural responses within the whisker region of the primary somatosensory cortex (SIw) of rats are profoundly influenced by the spatiotemporal attributes of ipsilateral, as well as contralateral, whisker stimuli. As inactivation of one SIw eliminates in the intact SIw both ipsilaterally evoked responses and the influence of ipsilateral stimulation on contralaterally evoked activity, we proposed that interhemispheric interactions between the SIws may be important for integrating bilateral whisker information. To test whether rats can recognize the bilateral nature of a whisker stimulus, we developed a tactile discrimination task that required rats to conjointly determine distances to a left and a right discriminandum as equidistant or non-equidistant using only their facial whiskers. All rats trained in this task achieved performance levels indicative of an ability to integrate bilateral whisker information. Testing during unilateral, as well as bilateral, inactivation of the SIws indicated that rats rely on both SIws for detecting the bilateral nature of a whisker stimulus. Rats were unable to perform the task without both sets of whiskers, a fact that indicates that the whiskers (and not other modalities) were used to perform this task. The findings presented here indicate that rats can solve a task that requires the conjoint detection of left and right whisker-mediated distance information and implicate the SIws as central to this ability.

## Introduction

Since the work of Sperry and colleagues on split-brain subjects, the role of interhemispheric interactions has been the object of intense investigation (Hamilton, 1998). Generally, the function of such interactions has been related to the ability to transfer a learned association between the hemispheres (Ebner and Myers, 1962). Additionally, interhemispheric interactions may mediate the ability to correlate images in the left and right visual hemifields, to join visual stimuli along the vertical meridian (as proposed in the ‘midline fusion hypothesis' (Chourdjury et al., 1965; Berlucchi et al., 1967; Hubel and Wiesel, 1967), to integrate somatosensation from paired limbs (Iwamura, 2000), or for unifying processes related to attention (Doty, 1995). Perhaps most importantly, as identified by Doty, interhemispheric interactions may be needed to yield a singularity of behavior ‘despite the potential for bi-hemispheric duplication of processes and experience' (Doty, 1995).

Sensory systems that are lateralized subcortically provide an ideal means of investigating the role of interhemispheric interactions in electrophysiological and behavioral studies. This anatomical feature is exemplified in the rat whisker system since pathways conveying whisker information are completely lateralized subcortically, finally converging in the whisker region of the primary somatosensory cortices (SIws) through roughly homotopic callosal projections (White and DeAmicis, 1977; Olavarria et al., 1984; Koralek et al., 1990; Cauller et al., 1998), but see Welker et al. (Welker et al., 1988) for non-homotopic connections. That such connections may provide an anatomical substrate for the integration of bilateral whisker information was recently investigated (Shuler et al., 2001). Not only were the SIws found to be responsive to ipsilateral whisker stimuli, in confirmation of Pidoux (Pidoux and Verley, 1979), but SIw responses were determined to be responsive in a manner dependent upon the spatial and temporal attributes of bilateral whisker stimuli (Shuler et al., 2001). In this study, we proposed that the mutual influence between each SIw provided the means by which rats could integrate left and right side whisker information, as might be required to determine the orientation of an obstacle or the width of an aperture.

Though a number of behavioral studies have examined rats' use of their whiskers to discriminate tactile features of the environment (Vincent, 1912; Welker, 1964; Gustafson and Felbain-Keramidas, 1977; Hutson and Masterton, 1986; Carvell and Simons, 1990; Brecht et al., 1997), no study to date has examined rats' ability to compare bilateral tactile features. However, a recent behavioral study does suggest that learned associations between unilateral whisker stimuli and reward are partially retained when stimuli are presented to homotopic (but not heterotopic) whiskers on the contralateral, untrained side of the face (Harris and Diamond, 2000). In keeping with findings from rat vision (Buresova and Nadel, 1970; Crowne et al., 1994), and other systems and species (Ettlinger and Elithorn, 1962; Mello, 1965; Noble, 1968), this intriguing result was interpreted as evidence that interhemispheric transfer occurs between homotopically interconnected regions of the SIws. This interpretation on the role of interhemispheric interactions raises the possibility, however, that the left and right components of a bilateral stimulus could be mutually antagonistic in their learned associations with obtaining reward. That such an antagonism could exist between the hemispheres was noted by Sperry, when reporting that an animal simultaneously trained in diametrically opposed unilateral tasks showed signs of interference when the corpus callosum was intact, whereas learning was unimpeded if the corpus callosum was cut (Sperry, 1964). Thus, if the role of interhemispheric interactions between the SIws of the rat is simply to transfer learned associations between the hemispheres, then one might expect that rats could not learn to associate bilateral whisker stimuli with reward if the left and right components of the stimulus were related to conflicting reward-seeking behaviors.

To test the experimental hypothesis that rats can integrate bilateral tactile stimuli, we developed a discrimination task in which rats learned to compare the relative distance of two ‘walls’, one to each side of the face, using only the facial whiskers. Thus, the task was rendered unambiguous only when the two distances were considered conjointly. We conceptualize such a task as a discrimination between equidistant and nonequidistant stimulus configurations (though no claim is made that rats use these constructs). The results presented here provide the first evidence that rats can indeed integrate bilateral whisker information. Further, we demonstrated using unilateral and bilateral reversible inactivation that the ability to integrate bilateral whisker stimuli is strongly dependent on both SIws being intact.

## Materials and Methods

To determine whether rats can integrate bilateral whisker information, we developed a novel bilateral tactile discrimination task. The behavioral apparatus was constructed to test the ability of rats to make tactile discriminations related to distance using only the large, facial whiskers of the whisker pad (Krupa et al., 2001), which we refer to as the vibrissal whiskers (or simply the whiskers), and to eliminate the use of other sensory cues. The behavioral apparatus of Krupa et al. (Krupa et al., 2001) was used as rats in that study were shown to quickly execute a similar whisker-mediated discrimination in a highly stereotypical manner.

### Behavioral Apparatus and Tactile Discrimination Task

The rat was required to simultaneously sample the position of the left and right discriminandum with its vibrissal whiskers prior to poking with its nose the cNP in the inner chamber, and to indicate the position of those discriminanda by poking either the reward nose-poke (lNP or rNP) in the outer chamber. The requirement that an infrared (IR) beam within the cNP be broken stereotyped the approach of the rat into the inner chamber. Rats learned to associate bilaterally equidistant or bilaterally non-equidistant whisker stimuli with the lNP and rNP, respectively. By breaking an IR beam within the lNP, for example, an equidistant stimulus configuration was indicated, whereas breaking an IR beam in the rNP indicated a non-equidistant stimulus configuration. If correct, a water reward (50 μl) was then delivered into the nose-poke (NP); if incorrect, no water was delivered.

Behavioral sessions were fully automated using the computer program MED-PC (Med Associates Inc., St Albans, VT), which logged and controlled the input and output of all sensors and effectors of the behavioral apparatus. Time-stamped behavioral events, as well as signals indicating trial type were sent to a second computer, a Multi-channel Acquisition Processor (MAP; Plexon, Inc., Dallas, TX). Finally, a video camera was used to record, under IR illumination, the movements of the animal during a session.

A number of requisite initial conditions were met prior to the start of a session. First, the animal was placed into the outer chamber with the center door closed blocking access to the inner chamber. The doors blocking access to the reward NPs were initially closed as well. The behavioral apparatus was then shut and, in turn, enclosed within a light-proof, sound-attenuating box. The left and right discriminanda were then positioned into their starting locations. Lastly, the behavioral apparatus, MAP and video recorder were activated.

### Behavioral Training

Eight Long Evans male rats weighing ~350 g at the start of training were used in this study. Rats were motivated to seek water rewards through performance in the behavioral task by placing them under a mild regimen of water restriction. Free access to water was limited to 30 min at the end of each day; however, food was available ad libitum. Though rats were trained on a daily basis, approximately every tenth day, rats were given free access to water for 24 h. Subsequent testing resumed after 2 days of water restriction.

Behavioral training proceeded through three phases. The first, and part of the second training phase, were conducted under incandescent house lights, whereas the latter portion of phase two and the third phase were conducted under IR illumination to preclude the use of visual cues. In the first phase of training, rats were confined to the outer chamber. By randomly opening the lNP or rNP door, access to the open reward NP was permitted. Rats quickly learned to associate poking their nose into the opened NP with water reward. After breaking the IR beam with their nose and receiving water, 10 s were allowed to elapse to permit rats to withdraw from the reward NP before the door was closed. This process was repeated with an inter-trial interval of 30 s. After a few days of training rats were moved to the second behavioral training phase.

The second phase of training introduced the association between breaking the cNP IR beam and opening a reward NP door in the outer chamber. Prior to the beginning of a trial, the left and right discriminanda were each placed into either near or far positions depending on the trial type that was randomly selected. The center door was then opened. Rats learned to enter into the inner chamber and poke their nose into the cNP. Breaking the cNP IR beam caused one of the two reward NP doors to open. If the trial was an equidistant trial type, the lNP door was opened; if the trial type was non-equidistant, the rNP door was opened. After breaking the cNP beam, rats quickly backed out of the inner chamber and sought out the open reward NP. Upon breaking the opened reward NP IR beam, the center door was closed, which was followed 10 s later by closure of the opened reward NP. This process continued with an inter-trial interval of 40 s (i.e. the time between when the center door was closed and when it was next opened).

After rats acquired a proficiency in the second phase, they were then moved to the third behavioral training phase. This phase differed from the second phase in one crucial aspect: breaking the cNP beam opened both reward NP doors. Therefore, in the third phase of training, rats had to learn to associate equidistant trial types with reward at the lNP and non-equidistant trial types with reward at the rNP. Breaking either NP ended the trial by closing the center door, but incorrect choice of NP resulted in no water reward. If, on a given trial, the rat failed to correctly associate the stimulus configuration with the appropriate reward NP, subsequent trials presented the same stimulus configuration until the rat performed the correct association. Such trials were termed correction trials and were excluded from analyses. A maximum of three correction trials were given before a forced trial was presented that required the correct NP IR beam to be broken for reward before testing could resume. Trials that did not follow those that were incorrectly answered were termed test trials.

### Stimulus Configurations Used

Since the left and right walls used as discriminanda in this task could be independently placed in either near or far positions, four possible stimulus configurations could be obtained, two equidistant and two non-equidistant. Specifically, the equidistant configurations of the discriminanda were (i) left far and right far (Wide) and (ii) left near and right near (Narrow), whereas the non-equidistant configurations were (iii) left far and right near (Leftwide) and (iv) left near and right far (Rightwide).

In the initial sessions of the third behavioral training phase, near and far discriminandum positions were set at 23 mm and 45 mm from the midline of the cNP, respectively. The absolute difference in near and far positions of a discriminandum was incrementally reduced through subsequent training until distances of 27.3 mm and 38.1 mm from the midline were obtained. Rats were trained until performance of the task reached asymptote. Having been well trained in the task, further manipulations to test the involvement of the SIws, as well as the whiskers themselves, were subsequently carried out.

### Implantation of Guide Cannula/Electrode Assembly into the SIws

Six trained animals were bilaterally implanted with fine caliber infusion guide cannulae attached to arrays of eight drivable electrodes to allow muscimol or saline to be injected into the whisker region of the SIws. For these rats, having once demonstrated a consistent proficiency in the discrimination task, water restriction and training was suspended. A week later, using stereotaxic measurements (–3.0 mm caudal from bregma, 5.5 mm mediolateral) cannula–electrode assemblies were implanted bilaterally so that injection cannulae tips would be ~1200 μm deep. Recording electrode arrays were implanted into superficial layers of the SIw at the time of surgery and were incrementally driven more deeply post surgery. At the time of surgery, the electrode arrays permitted neural activity evoked by stimulation of contralateral whiskers to be monitored via audiomonitor. Buprenophrine (0.1–0.5 mg/kg, s.c.) was used for post-surgical analgesia and a topical antibiotic was applied along the implanted assembly. Following 7 days of post-surgical recovery, rats were once again placed on water restriction, whereupon training was again resumed.

### Experimental Manipulations

After rats consistently performed the task and showed no further signs of improvement, a series of experimental manipulations (along with their controls) were conducted. The manipulations included (i) inactivating the left SIw; (ii) inactivating the right SIw; (iii) inactivating the left and right SIws; (iv) cutting off all the vibrissal whiskers unilaterally; and (v) cutting off all the whiskers bilaterally. Individual inactivation of the left or right SIw was conducted to address what extent the use of bilateral whisker information would be diminished. Inactivating both SIws was used to examine the extent performance of the task was dependent on the SIws. To control for any apparent ability to use bilateral whisker information regarding the position of the left and right discriminandum, rats were tested after unilaterally cutting the vibrissae from one side of the face. Finally, cutting off the whiskers bilaterally was itself a control to indicate that rats used their whiskers, and not other means, to determine the positions of the discriminanda. Performance in the task was recorded for the day of each manipulation as well as for two consecutive days immediately prior to each manipulation. Following each manipulation, additional sessions were needed to again establish that performance levels were stable.

In the case of manipulations using muscimol, saline was infused as a control 1 day before muscimol inactivation. In rats lightly anesthetized with halothane (2% in oxygen), 500 nl of saline was slowly injected (100 nl/min) into either the left, right or both SIws an hour prior to the session by using a microperfusion pump (Orien Research, Inc., Beverly, MA). The next day, muscimol (500 ng in 500 nl of saline) was injected using the same procedure. This volume and concentration of muscimol was determined to be adequate to inactivate a large portion of the SIw (Martin, 1991; Krupa et al., 1999; Shuler et al., 2001). After completion of the infusion, a minimum of 45 min was allowed to elapse before the beginning of a session. Muscimol inactivations were confirmed by the absence of neural activity from electrodes implanted into the SIw prior to, and shortly after, the completion of the session. In all cases, neural activity was absent in the inactivated SIw.

The fourth experimental manipulation was that of removing the vibrissal whiskers on one side of the face. For this manipulation, rats were lightly anesthetized under halothane 14–16 h prior to testing and the whiskers were cut unilaterally leaving only the small whiskers around the nose and lips, as well as the fur, intact. Following testing, the other side of the face was subsequently cut and 6 weeks were allowed to elapse to permit the whiskers to regrow on each side of the face. During this time the rats had free access to water and food and all testing was suspended. After the whiskers had fully regrown on both sides of the face, rats were placed back on water restriction and again tested in the task. After stable performance levels were established, three consecutive days of performance were recorded. As in the previous manipulation, after the second day, rats were anesthetized so as to cut the whiskers off bilaterally.

In addition to measuring performance, the total number of trials in a session before and after manipulation was also recorded to provide a general measure of motivation. Together with off-line analysis of video tape recordings, the impact of experimental manipulations on rats' general motivation and ability to perform the mechanics of the task was assessed.

### Performance Space and Analysis

The intent of this experiment was to determine if rats could integrate bilateral whisker information. To test this possibility, a behavioral paradigm was designed such that reward could be obtained at its highest rate only if information regarding the position of both left and right discriminandum were taken into account.

If, in a hypothetical session, a rat performed a total of 76 test trials with Narrow and Wide stimulus configurations as the equidistant stimuli, and Leftwide as the non-equidistant stimulus, 19 Narrow, 19 Wide and 38 Leftwide test trials would have been given. As displayed in Table 1, had the hypothetical performance consisted of 17 Narrow correct, 17 Wide correct and 34 Leftwide correct, the percent correct for Narrow, Wide and Leftwide trials would each have been 89%.

### A Metric for the Degree of Bilateral Whisker Use

$2\ {\times}\ {[}(1\ {\mbox{--}}\ (distance\ to\ Bilateral\ Best\ strategy/hypotenuse\ of\ performance\ space))\ {\mbox{--}}\ 0.5{]}$
The distance of the observed session performance from the Bilateral Best strategy, taken as a proportion of the distance between the Bilateral Best and Bilateral Worst strategies (the hypotenuse of the strategy space) is subtracted from unity. This fraction is then subtracted by 0.5 so that the mean of possible values lies at zero, and is then multiplied by 2 to scale the range of potential index values from –1 to 1. A value of 1 then corresponds to absolute certainty that the animal performed the task by perfectly associating the bilateral distance information with the correct reward NPs, whereas a value of –1 corresponds to absolute certainty that the animal perfectly associated bilateral distance information with the incorrect reward NPs. Finally, a value of zero would indicate that the animal could not integrate bilateral whisker information.

### Histology

The placement of the cannula–electrode assemblies into the SIws was confirmed histologically (Fig. 4A,B) using a staining procedure for cytochrome oxidase adapted from Divac (Divac et al., 1995). All cannulae used in inactivations were localized to the SIw. Some cannulae could not be used due to obstruction. Rats were killed with a lethal dose of sodium pentobarbital. The brain was removed, and the left and right cerebral cortices were separated from underlying tissue, placed flat within a container, and then were quickly frozen. Parasagittal, 20 μm sections were then cut through the entire depth of the cortex with a cryostat. Sections mounted on slides were then reacted for cytochrome oxidase and then coverslipped.

## Results

### General Observations

During execution of the task rats adopted a highly stereotyped approach/retreat into the inner discrimination chamber that typically lasted ~1 s (see Fig. 2). As depicted in Figure 2, this process can be described as a sequence of epochs: (i) Approach, (ii) Contact, (iii) Center Poke, (iv) Decision Made and (v) Retreat. For all animals observed under all conditions, immediately upon opening the center door, rats were observed to quickly approach the cNP in the inner chamber. Though rats were never observed to make ‘whisking’ movements of the whiskers on approach to the cNP, a subtle protraction of the whiskers was observed. In combination with the animals' forward movement, this whisker protraction resulted in a backward deflection of the whiskers as they struck the discriminandum. Rats were observed to make contact with the left and right discriminandum with only their left and right vibrissal whiskers, respectively. The median time between first contacting the discriminanda and breaking the cNP was 280 ms and the inter-quartile range was 222–357 ms (taken from all eight animals for 1 day of testing). Rats sampled the discriminanda only once per trial. Retracting their nose from the cNP, rats were observed to turn their head, as they retreated from the inner chamber, slightly to the side that corresponded to the NP in which they then sought reward. In accord with a similar testing paradigm on which this experiment was based (Krupa et al., 2001), rats were not observed to use alternative movements, such as making side-to-side head movements or contacting the discriminanda with other body parts to perform the task.

### Comparisons Between Trained and Experimentally Manipulated Groups

Outside of a diminished ability to correctly perform the task, videotape analysis did not reveal any overt differences in behavior following any experimental manipulation. As a measure of general motor ability, an analysis of the speed at which rats entered into the inner chamber also failed to determine any significant difference following experimental manipulation. Experimental manipulation only minimally affected the total number of trials executed in a session compared to control levels (88 versus 91 trials), supporting the proposition that experimental manipulations did little to alter motivational state.

No significant differences were found for the percent of correctly answered trials or for the BUI values between the two control days prior to testing for each of the experimental manipulation groups (including the normal, trained group). A 6 × 2 mixed-effect ANOVA, with experimental manipulation as the six-level between-group effect, and condition (prevs postmanipulation) as the two-level effect, confirmed the existence of significant differences in the percent of correctly answered trials between the groups [F(5,24) = 10.52; P < 0.0001]. The Trained group differed from all experimentally manipulated groups in that percent correct did not significantly change across the days of testing, whereas a significant diminishment in percent correct, preversus post-manipulation, occurred in all other groups. The Bilateral Cut group resulted in a significantly greater decrement than all other manipulations (LSD post hocs; P values ≤ 0.038), which were indistinguishable from each other.

A 6 × 2 mixed-effect ANOVA, like that performed for percent of correctly answered trials, was also conducted for BUI values and confirmed the existence of significant differences between the groups [F(5,24) = 11.55; P < 0.0005]. Subsequent post hoc tests determined that the Trained group was significantly different than all experimental manipulation groups (P values ≤ 0.001) as no decrement in bilateral integration in the Trained group occurred across the days of testing. This result was in contrast to all other groups in that any experimental manipulation caused a significant decrement in BUI values. The Bilateral Cut group resulted in a greater decrement in BUI values than either Left Inactivation (P = 0.04) or Right Inactivation (P = 0.01), though it was not distinguishable from Bilateral Inactivation or Unilateral Cut.

## Discussion

Trained animals executed the task at performance levels consistent only with the ability to regard the conjoint positions of the left and the right discriminandum. This result confirms that a fundamental ability of the whisker system is to detect the distance to objects in the environment, as proposed by Welker et al. (Welker et al., 1988) and as formalized in the distance detection hypothesis by Brecht et al. (Brecht et al., 1997). Though behavioral studies have implicated the whisker system in the ability to detect the presence/absence of objects (Hutson and Masterton, 1986; Brecht et al., 1997), the notion that whiskers are used to detect distances was demonstrated only recently by work from our laboratory (Krupa et al., 2001). Though this recent work suggests that the whiskers do not function as simple binary on/off elements as suggested by Brecht et al. (Brecht et al., 1997), our results support their view that the rat vibrissal system detects ‘head-centered' contours of obstacles or openings. The present study provides the first experimental evidence that distance detection can be synthesized bilaterally, as would be required to detect the orientation of an obstacle or width of an aperture. As information from both whisker pads is a requirement of the task, we also provide strong evidence that the rat whisker system integrates multi-whisker input. Further, in systems where sensory input is lateralized subcortically, few explicit tests of the ability to compare stimuli between the hemispheres exist, and deal primarily with the visual system (Hamilton et al., 1973; Lewine et al., 1994). In a broader context, then, this finding provides behavioral evidence in somesthesis that bihemispheric compound stimuli can be discriminated.

Individual inactivation of the left or right SIw significantly diminished the percent of correctly answered trials as well as the ability to conjointly use leftand right-side stimulus-related information. As might have been expected considering electrophysiological evidence demonstrating bilateral interactions between the SIws, unilateral inactivation did not equivalently disrupt the ability of rats to associate each of the three stimulus configurations with reward. Left SIw inactivation, for instance, resulted in an inability to associate with reward either the Wide or Leftwide configurations, but an ability to correctly associate the Narrow configuration remained. This result is curious in that performance levels obtained are not consistent with a unilateral strategy that would have optimized performance if information from only one side of the face could have been used (see S3, Unilateral Left Best or S4, Unilateral Right Best). A potential explanation is put forth in Figure 8 to account for the results obtained during unilateral inactivation. If, as postulated, and supported by Hutson and Masterton (Hutson and Masterton, 1986), SIw removal (or in this case SIw inactivation) is tantamount to removing the contralateral whiskers themselves, unilateral SIw inactivation effectively occludes the contralateral tactile hemifield. As the rat could then rely only on the whiskers contralateral to the ‘intact’SIw, only one of the three potential positions of the discriminanda could be unambiguously associated with reward. Therefore, when the left SIw is inactivated, the Narrow condition was unilaterally unambiguous using the left whiskers, whereas when the right SIw was inactivated, the Wide condition was unilaterally unambiguous using the right whiskers. Though this accounts for the behavior observed across animals during unilateral inactivation (left or right), this explanation requires that a stimulus never before associated with any particular behavior was generalized to a behavior that ensured reward. When the bias to the unilaterally unambiguous stimulus is taken into account, no significant difference could be determined between left inactivation and right inactivation. These results strongly indicate that both SIws are involved in performance of the task.

A severe impairment in the ability to perform the discrimination followed bilateral inactivation of the SIws. Though the percent of correct trials was slightly above chance following bilateral inactivation, we believe that the most parsimonious explanation for any residual ability following bilateral inactivation may have resulted from an inability to completely inactivate the SIws. However, the involvement of other structures cannot be fully ruled out. Nonetheless, the degree to which unilateral and bilateral inactivation diminished the ability to regard bilateral whisker information were indistinguishable as predicted if rats truly used both SIws to process bilateral whisker information to perform the task.

The apparent ability of rats to integrate bilateral whisker information was controlled for by testing performance in rats that had undergone unilateral whisker removal. Removing the whiskers unilaterally prior to testing resulted in a profound diminishment in the percent of correctly answered trials. As expected, unilateral whisker removal precluded the ability of rats to correctly associate the bilateral nature of the stimulus with reward, confirming that rats did use their whiskers to obtain bilateral tactile information. Similarly to left or right inactivations, whisker removal on one side of the face biased performance such that the position of the unilaterally unambiguous discriminandum for the whisker-intact side of the face was associated with reward significantly above chance levels. Taken together, these results strongly indicate that rats did use their whiskers to obtain bilateral tactile information used to perform the task.

In regards to the potential of unilateral whisker removal to have resulted in a bias toward one side of the chamber (thigmotaxis) as previously demonstrated (Milani et al., 1989), there was no systematic difference in side preference preversus post-removal that related to the whisker-intact side of the face or, for that matter, to any of the experimental manipulations. Considering that in prior reports, a bias toward using the whisker-intact side of the face was no longer detected after moderate exposure to the testing environment, perhaps no bias was to be expected here, in that the rats tested were well trained and accustomed to the behavior chamber.

Lastly, bilateral whisker removal resulted in a drop in the percent of correctly answered trials and a drop in the degree to which rats usefully employed leftand right-side information that was indistinguishable from the Guessing strategy. This control manipulation provided strong confirmation that rats use whisker-mediated tactile information to perform the task.

The current study provides behavioral evidence that rats can in fact integrate bilateral whisker information, and that both whisker barrel cortices are required. If the role of interhemispheric interactions between the SIws is to transfer learned associations in one SIw to its mirror-reflected counterpart in the opposite SIw, then the finding that rats can solve a bilateral task where unilateral components are antagonistically related with reward should be viewed as unexpected. However, this viewpoint stems from the supposition that the SIws are separate entities; one processing left whisker information and one processing right whisker information. Alternatively, if the SIws are viewed as one system, then interhemispheric interactions may serve as the means by which left and right whisker information is integrated to yield unified bilateral tactile percepts. As noted in the Methods, however, rats were apparently incapable of associating all four possible stimulus configurations with reward. As such a task requires exclusive–OR logic, perhaps the inability of rats to perform the task suggests a psychophysical limitation related to how interhemispheric interactions permit bi-hemispheric associations with reward to be made. Having firmly established that rats can integrate bilateral whisker information and that the SIws are needed, further behavioral experiments combining electrophysiological recordings with reversible callosal inactivation will be necessary to address the role of interhemispheric interactions between the SIws.

## Notes

This research was supported by a grant from the National Institute of Dental Research Grants DE11121-01 and DE13810-01 to M.A.L.N. and an NRSA grant MH12570-01A1 to M.G.S. We thank Ben Rubin, Don Katz, Ford Ebner and David Fitzpatrick, for their helpful insights and criticisms.

Table 1

Comparison of example session performance against idealized strategies: % of correctly answered test trials by stimulus configuration

Equidistant Non-equidistant
Strategies Narrow (xWide (yLeftwide (zOverall % correct
Session performance is given for each of nine idealized strategies for the expected percent of correctly answered Narrow (equidistant) and Wide (equidistant), and Leftwide (non-equidistant) test trials, as well as the percent correct, in toto. Bilateral Best (S1) indicates a perfect association between the bilateral configuration of the stimulus and reward NP. Conversely, Bilateral Worst (S2) indicates that the bilateral configuration of the stimulus was perfectly matched with the wrong NPs. Unilateral Left Best (S3) and Unilateral Right Best (S4), relate the best performance obtainable if rats were limited to using only unilateral whisker information. Conversely, if rats were limited to unilateral whisker information yet failed to ensure association of the most frequent stimulus configuration with reward, then only 25% of the trials would be answered correctly as in Unilateral Left Worst (S5), and Unilateral Right Worst (S6) strategies. Left Bias (S7) and Right Bias (S8) correspond to rats eschewing the cues provided by the position of the discriminanda for perfect biases to the equidistant NP or to the non-equidistant NP, respectively. Finally, if rats were either incapable of perceiving the positions of the stimuli, learning their associated NPs, or were indifferent to maximizing reward, a strategy whereby rats simply guessed at the correct NP is proposed, Guessing (S9). A hypothetical example session that consisted of 76 test trials is also provided for comparison, such that 17/19 Narrow, 17/19 Wide and 34/38 Leftwide trials were answered correctly (89% for each). Percent correct of Narrow, Wide and Leftwide trials provide the x, y and z coordinates for plotting these performances in the performance space of the accompanying Figure 3
S1 Bilateral Best 100 100 100 100
S2 Bilateral Worst
S3Unilateral Left Best 100 100 75
S4Unilateral Right Best 100 100 75
S5Unilateral Left Worst 100 25
S6Unilateral Right Worst 100 25
S7Left Bias 100 100 50
S8Right Bias 100  50
S9 Guessing 50 50 50 50
Hypothetical session 89 89 89 89
Equidistant Non-equidistant
Strategies Narrow (xWide (yLeftwide (zOverall % correct
Session performance is given for each of nine idealized strategies for the expected percent of correctly answered Narrow (equidistant) and Wide (equidistant), and Leftwide (non-equidistant) test trials, as well as the percent correct, in toto. Bilateral Best (S1) indicates a perfect association between the bilateral configuration of the stimulus and reward NP. Conversely, Bilateral Worst (S2) indicates that the bilateral configuration of the stimulus was perfectly matched with the wrong NPs. Unilateral Left Best (S3) and Unilateral Right Best (S4), relate the best performance obtainable if rats were limited to using only unilateral whisker information. Conversely, if rats were limited to unilateral whisker information yet failed to ensure association of the most frequent stimulus configuration with reward, then only 25% of the trials would be answered correctly as in Unilateral Left Worst (S5), and Unilateral Right Worst (S6) strategies. Left Bias (S7) and Right Bias (S8) correspond to rats eschewing the cues provided by the position of the discriminanda for perfect biases to the equidistant NP or to the non-equidistant NP, respectively. Finally, if rats were either incapable of perceiving the positions of the stimuli, learning their associated NPs, or were indifferent to maximizing reward, a strategy whereby rats simply guessed at the correct NP is proposed, Guessing (S9). A hypothetical example session that consisted of 76 test trials is also provided for comparison, such that 17/19 Narrow, 17/19 Wide and 34/38 Leftwide trials were answered correctly (89% for each). Percent correct of Narrow, Wide and Leftwide trials provide the x, y and z coordinates for plotting these performances in the performance space of the accompanying Figure 3
S1 Bilateral Best 100 100 100 100
S2 Bilateral Worst
S3Unilateral Left Best 100 100 75
S4Unilateral Right Best 100 100 75
S5Unilateral Left Worst 100 25
S6Unilateral Right Worst 100 25
S7Left Bias 100 100 50
S8Right Bias 100  50
S9 Guessing 50 50 50 50
Hypothetical session 89 89 89 89
Table 2

Summary by group of bilateral whisker discrimination performance

Experimental group Condition Total trials Test trials Correct test trials Incorrect test trials % correct test trials
Each experimental group is divided into control and testing days. The average number of total trials, test trials, correct test trials, incorrect test trials, and percent correct are given along with the standard error of the means in parentheses.
Trained day 1 90.75 (0.94) 78.37 (2.03) 68.12 (3.00) 10.25 (1.21) 86.67 (1.71)
day 2 88.87 (1.08) 75.38 (2.06) 65.37 (3.13) 10.00 (1.28) 86.42 (1.96)
day 3 89.87 (1.17) 77.75 (2.00) 67.50 (3.02) 10.25 (1.32) 86.57 (1.83)
day 4 90.25 (1.06) 78.00 (1.92) 68.00 (2.68) 10.00 (0.94) 86.97 (1.41)
day 5 89.75 (1.77) 75.12 (3.04) 64.37 (4.40) 10.75 (1.66) 85.06 (2.56)
Left inactivation pre-saline 91.75 (0.48) 77.50 (0.96) 65.25 (1.38) 12.25 (0.48) 84.16 (0.79)
saline 91.00 (0.71) 77.75 (1.60) 66.75 (2.87) 11.00 (1.47) 85.74 (2.14)
inactivation 86.50 (1.85) 58.50 (1.94) 39.00 (2.16) 19.50 (1.32) 66.58 (2.46)
Right inactivation pre-saline 91.40 (0.68) 78.20 (0.97) 66.80 (1.39) 11.40 (0.68) 85.39 (0.93)
saline 91.40 (0.75) 78.00 (1.38) 67.00 (2.05) 11.00 (0.77) 85.82 (1.24)
inactivation 82.60 (4.48) 55.80 (2.94) 37.80 (2.42) 18.00 (1.14) 67.65 (1.73)
Bilateral inactivation pre-saline 92.00 (0.58) 80.00 (4.16) 70.67 (5.24)  9.33 (1.20) 88.12 (2.07)
saline 91.33 (0.88) 78.33 (3.28) 68.33 (4.98) 10.00 (1.73) 87.00 (2.79)
inactivation 90.00 (0.58) 53.33 (2.73) 32.00 (1.53) 21.33 (1.45) 60.04 (1.15)
Unilateral cut pre, pre-cut 88.80 (3.50) 78.40 (4.65) 69.80 (5.42) 8.60 (1.03) 88.58 (2.05)
pre-cut 91.60 (0.60) 80.00 (2.35) 69.60 (4.23) 10.40 (1.94) 86.70 (2.64)
cut 87.60 (2.36) 51.00 (1.67) 30.20 (1.56) 20.80 (0.37) 59.08 (1.18)
Bilateral cut pre, pre-cut 89.20 (1.24) 73.20 (1.56) 61.60 (2.09) 11.60 (0.81) 84.07 (1.48)
pre-cut 90.40 (1.08) 76.80 (2.01) 65.40 (2.71) 11.40 (0.87) 85.02 (1.48)
cut 87.00 (1.64) 50.40 (4.03) 25.40 (2.94) 25.00 (2.12) 50.12 (3.30)
Experimental group Condition Total trials Test trials Correct test trials Incorrect test trials % correct test trials
Each experimental group is divided into control and testing days. The average number of total trials, test trials, correct test trials, incorrect test trials, and percent correct are given along with the standard error of the means in parentheses.
Trained day 1 90.75 (0.94) 78.37 (2.03) 68.12 (3.00) 10.25 (1.21) 86.67 (1.71)
day 2 88.87 (1.08) 75.38 (2.06) 65.37 (3.13) 10.00 (1.28) 86.42 (1.96)
day 3 89.87 (1.17) 77.75 (2.00) 67.50 (3.02) 10.25 (1.32) 86.57 (1.83)
day 4 90.25 (1.06) 78.00 (1.92) 68.00 (2.68) 10.00 (0.94) 86.97 (1.41)
day 5 89.75 (1.77) 75.12 (3.04) 64.37 (4.40) 10.75 (1.66) 85.06 (2.56)
Left inactivation pre-saline 91.75 (0.48) 77.50 (0.96) 65.25 (1.38) 12.25 (0.48) 84.16 (0.79)
saline 91.00 (0.71) 77.75 (1.60) 66.75 (2.87) 11.00 (1.47) 85.74 (2.14)
inactivation 86.50 (1.85) 58.50 (1.94) 39.00 (2.16) 19.50 (1.32) 66.58 (2.46)
Right inactivation pre-saline 91.40 (0.68) 78.20 (0.97) 66.80 (1.39) 11.40 (0.68) 85.39 (0.93)
saline 91.40 (0.75) 78.00 (1.38) 67.00 (2.05) 11.00 (0.77) 85.82 (1.24)
inactivation 82.60 (4.48) 55.80 (2.94) 37.80 (2.42) 18.00 (1.14) 67.65 (1.73)
Bilateral inactivation pre-saline 92.00 (0.58) 80.00 (4.16) 70.67 (5.24)  9.33 (1.20) 88.12 (2.07)
saline 91.33 (0.88) 78.33 (3.28) 68.33 (4.98) 10.00 (1.73) 87.00 (2.79)
inactivation 90.00 (0.58) 53.33 (2.73) 32.00 (1.53) 21.33 (1.45) 60.04 (1.15)
Unilateral cut pre, pre-cut 88.80 (3.50) 78.40 (4.65) 69.80 (5.42) 8.60 (1.03) 88.58 (2.05)
pre-cut 91.60 (0.60) 80.00 (2.35) 69.60 (4.23) 10.40 (1.94) 86.70 (2.64)
cut 87.60 (2.36) 51.00 (1.67) 30.20 (1.56) 20.80 (0.37) 59.08 (1.18)
Bilateral cut pre, pre-cut 89.20 (1.24) 73.20 (1.56) 61.60 (2.09) 11.60 (0.81) 84.07 (1.48)
pre-cut 90.40 (1.08) 76.80 (2.01) 65.40 (2.71) 11.40 (0.87) 85.02 (1.48)
cut 87.00 (1.64) 50.40 (4.03) 25.40 (2.94) 25.00 (2.12) 50.12 (3.30)
Figure 1.

Schematic of the behavioral apparatus (A), and flow chart of the bilateral whisker discrimination task (B).

Figure 1.

Schematic of the behavioral apparatus (A), and flow chart of the bilateral whisker discrimination task (B).

Figure 2.

Discrimination by a trained rat of Narrow and Wide (equidistant), and Leftwide (non-equidistant) stimulus configurations. The stereotyped approach of the rat is decomposed into five epochs: Approach, Contact, Center Poke, Decision Made, and Retreat. Black scale bars = 27.3 mm distance to near discriminandum position. White scale bars = 38.1 mm distance to far discriminandum position. In the Decision Made and Retreat epochs, the rat can be seen to be turning in the direction of the correct NP for reward as it retreats from the inner chamber. The non-equidistant stimulus configuration (Leftwide) is associated with the right reward NP, whereas the equidistant stimulus configurations (Narrow and Wide) are associated with the left reward NP. Timestamps in minutes, seconds, centiseconds.

Figure 2.

Discrimination by a trained rat of Narrow and Wide (equidistant), and Leftwide (non-equidistant) stimulus configurations. The stereotyped approach of the rat is decomposed into five epochs: Approach, Contact, Center Poke, Decision Made, and Retreat. Black scale bars = 27.3 mm distance to near discriminandum position. White scale bars = 38.1 mm distance to far discriminandum position. In the Decision Made and Retreat epochs, the rat can be seen to be turning in the direction of the correct NP for reward as it retreats from the inner chamber. The non-equidistant stimulus configuration (Leftwide) is associated with the right reward NP, whereas the equidistant stimulus configurations (Narrow and Wide) are associated with the left reward NP. Timestamps in minutes, seconds, centiseconds.

Figure 4.

Histology and schematic of the primary somatosensory cortex. (A) 20 μ tangential cytochrome oxidase stained section of the primary somatosensory cortex. Lines demarcate the whisker region (SIw) corresponding to the large vibrissal whiskers. Closed arrow marks the position of the injector cannula, approximately in the C2 whisker barrel. Electrode lesions from the implanted electrode array marked by open arrows are found along the caudomedial border of the SIw. (B) The positions of the cannulae used in animals that resulted in successful infusion of saline/muscimol are superimposed on a schematic of the SIw.

Figure 4.

Histology and schematic of the primary somatosensory cortex. (A) 20 μ tangential cytochrome oxidase stained section of the primary somatosensory cortex. Lines demarcate the whisker region (SIw) corresponding to the large vibrissal whiskers. Closed arrow marks the position of the injector cannula, approximately in the C2 whisker barrel. Electrode lesions from the implanted electrode array marked by open arrows are found along the caudomedial border of the SIw. (B) The positions of the cannulae used in animals that resulted in successful infusion of saline/muscimol are superimposed on a schematic of the SIw.

Figure 5.

Percent of correctly answered test trials ± SEM. (A) Five consecutive days of testing for trained animals. (B) Percent of correct test trials pre-saline, saline, and during left SIw inactivation. (C) Percent of correct test trials pre-saline, saline, and during right SIw inactivation. (D) Percent of correct test trials pre-saline, saline, and during bilateral SIw inactivation. (E) Percent of correct test trials for 2 days prior to and the day of testing with whiskers removed unilaterally. (F) Percent of correct test trials for 2 days prior to and the day of testing with whiskers removed bilaterally.

Figure 5.

Percent of correctly answered test trials ± SEM. (A) Five consecutive days of testing for trained animals. (B) Percent of correct test trials pre-saline, saline, and during left SIw inactivation. (C) Percent of correct test trials pre-saline, saline, and during right SIw inactivation. (D) Percent of correct test trials pre-saline, saline, and during bilateral SIw inactivation. (E) Percent of correct test trials for 2 days prior to and the day of testing with whiskers removed unilaterally. (F) Percent of correct test trials for 2 days prior to and the day of testing with whiskers removed bilaterally.

Figure 6.

Session performance plotted in performance space for trained animals under normal and manipulated circumstances. X-axis: % of correctly answered Narrow test trials. Y-axis: % of correctly answered Wide test trials. Z-axis: % of correctly answered Leftwide test trials. Performance across animals is plotted such that the standard error of the mean along the x, y and z dimensions form an ellipsoid centered on the average percent correct for each of the three stimulus configurations that constituted a session. Guessing strategy (star). Bilateral Best strategy (circle). Partition – a plane representing points in space that are equidistant to the Guessing and Bilateral Best strategies. (A) Performance of trained animals for five consecutive days. Five ellipsoids, one for each day of testing, are nested into each other in a dense cluster that lies above the partition close to the Bilateral Best strategy. (B) Left SIw inactivation. (C) Right SIw inactivation. (D) Bilateral SIw inactivation. (E) Unilateral whisker removal. (F) Bilateral whisker removal. As can be seen for performance in the experimental manipulation groups, control days prior to manipulation lie closest to the Bilateral Best strategy, whereas performance during manipulation falls beneath the partition. The average x, y and z coordinates are given in parentheses following each experimental manipulation. For simplicity, the axes were transformed for rats trained to discriminate Narrow from Leftwide and Rightwide (Narrow = z-axis; Leftwide = x-axis; Rightwide = y-axis).

Figure 6.

Session performance plotted in performance space for trained animals under normal and manipulated circumstances. X-axis: % of correctly answered Narrow test trials. Y-axis: % of correctly answered Wide test trials. Z-axis: % of correctly answered Leftwide test trials. Performance across animals is plotted such that the standard error of the mean along the x, y and z dimensions form an ellipsoid centered on the average percent correct for each of the three stimulus configurations that constituted a session. Guessing strategy (star). Bilateral Best strategy (circle). Partition – a plane representing points in space that are equidistant to the Guessing and Bilateral Best strategies. (A) Performance of trained animals for five consecutive days. Five ellipsoids, one for each day of testing, are nested into each other in a dense cluster that lies above the partition close to the Bilateral Best strategy. (B) Left SIw inactivation. (C) Right SIw inactivation. (D) Bilateral SIw inactivation. (E) Unilateral whisker removal. (F) Bilateral whisker removal. As can be seen for performance in the experimental manipulation groups, control days prior to manipulation lie closest to the Bilateral Best strategy, whereas performance during manipulation falls beneath the partition. The average x, y and z coordinates are given in parentheses following each experimental manipulation. For simplicity, the axes were transformed for rats trained to discriminate Narrow from Leftwide and Rightwide (Narrow = z-axis; Leftwide = x-axis; Rightwide = y-axis).

Figure 7.

Bilateral Use Index (BUI) scores for all experimental groups. A significant difference existed between groups as confirmed by a 6×2 mixed-effect ANOVA, with group as the six-level between-group effect and preversus post-manipulation as the two-level within-group effect [F(5,24) = 11.55; P < 0.0005]. Though post hoc tests (LSD) determined no significant difference in BUI scores for tested days in trained animals, a significant decrement in BUI followed all experimental manipulations (P values ≤ 0.04). Though not distinguishable from Bilateral Inactivation or Unilateral Cut, bilateral whisker removal resulted in a greater decrement in BUI values than either Left Inactivation or Right Inactivation (P values ≤ 0.04).

Figure 7.

Bilateral Use Index (BUI) scores for all experimental groups. A significant difference existed between groups as confirmed by a 6×2 mixed-effect ANOVA, with group as the six-level between-group effect and preversus post-manipulation as the two-level within-group effect [F(5,24) = 11.55; P < 0.0005]. Though post hoc tests (LSD) determined no significant difference in BUI scores for tested days in trained animals, a significant decrement in BUI followed all experimental manipulations (P values ≤ 0.04). Though not distinguishable from Bilateral Inactivation or Unilateral Cut, bilateral whisker removal resulted in a greater decrement in BUI values than either Left Inactivation or Right Inactivation (P values ≤ 0.04).

Figure 8.

A potential explanation of observed performance following unilateral inactivation or unilateral whisker removal. Animals perform the task in accord with the fact that unilaterally, one of the three stimulus configurations can be unambiguously associated with reward. (A) If inactivation of the left SIw results in a complete occlusion of right whisker information, using only the left whiskers, the Narrow configuration could be unambiguously associated with reward, whereas the Wide and Leftwide configurations could not be unambiguously associated with reward. (B) Converse to the effect of left SIw inactivation, if the right SIw is inactivated, the Narrow and Leftwide stimulus configurations are ambiguous with regard to the right whiskers, whereas the Wide stimulus configuration can be unambiguously associated with the reward.

Figure 8.

A potential explanation of observed performance following unilateral inactivation or unilateral whisker removal. Animals perform the task in accord with the fact that unilaterally, one of the three stimulus configurations can be unambiguously associated with reward. (A) If inactivation of the left SIw results in a complete occlusion of right whisker information, using only the left whiskers, the Narrow configuration could be unambiguously associated with reward, whereas the Wide and Leftwide configurations could not be unambiguously associated with reward. (B) Converse to the effect of left SIw inactivation, if the right SIw is inactivated, the Narrow and Leftwide stimulus configurations are ambiguous with regard to the right whiskers, whereas the Wide stimulus configuration can be unambiguously associated with the reward.

## References

Berlucchi G, Gazzaniga MS, Rizzolatti G (
1967
) Microelectrode analysis of transfer of visual information by the corpus callosum.
Arch Ital Biol

105
:
583
–596.
Brecht M, Preilowski B, Merzenich M (
1997
) Functional architecture of the mystacial vibrissae.
Behav Brain Res

84
:
81
–97.
1970
) Interhemispheric transfer in the rat.
Physiol Behav

5
:
849
–853.
Carvell GE, Simons DJ (
1990
) Biometric analyses of vibrissal tactile discrimination in the rat.
J Neurosci

10
:
2638
–2648.
Cauller L, Clancy B, Connors B (
1998
) Backward cortical projections to primary somatosensory cortex in rats extend long horizontal axons in layer 1.
J Comp Neurol

390
:
297
–310.
Chourdjury BP, Whitteridge D, Wilson ME (
1965
) The function of the callosal connections of the visual cortex.
Q J Exp Psychol B

50
:
214
–219.
Crowne DP, Forsyth P, Fitzgerald J (
1994
) Interocular transfer in the split-brain rat.
Behav Neurosci

108
:
804
–809.
Divac I, Mojsiilovic-Petrovic J, Lopez-Figueroa M, Petrovic-Minic B, Moller M (
1995
) Improved contrast in histochemical detection of cytochrome oxidase: metallic ions protocol.
J Neurosci Methods

56
:
105
–113.
Doty RW (
1995
) The five mysteries of the mind, and their consequences.
Neuropsychologia

36
:
1069
–1076.
Ebner FF, Myers RE (
1962
) Corpus callosum and the interhemispheric transmission of tactual learning.
J Neurophysiol

25
:
380
–391.
Ettlinger G, Elithorn A (
1962
) Transfer between the hands of a mirrorimage tactile shape discrimination.
Nature

194
:
1101
.
Gustafson J, Felbain-Keramidas S (
1977
) Behavioral and neural approaches to the function of the mystacial vibrissae.
Psychol Bull

84
:
477
–488.
Hamilton CR (
1998
) Paths in the brain, actions of the mind: Special issue in honor of Roger W. Sperry.
Neuropsychologia

36
:
953
–954.
Hamilton CR, Tieman SB, Brody BA (
1973
) Interhemispheric comparison of mirror image stimuli in chiasm-sectioned monkeys.
Brain Res

58
:
415
–425.
Harris JA, Diamond ME (
2000
) Ipsilateral and contralateral transfer of tactile learning.
NeuroReport

11
:
263
–266.
Hubel DH, Wiesel TN (
1967
) Cortical and callosal connections concerned with the vertical meridian of visual fields in the cat.
J Neurophysiol

30
:
1561
–1573.
Hutson K, Masterton R (
1986
) The sensory contribution of a single vibrissa's cortical barrel.
J Neurophysiol

56
:
1196
–1123.
Iwamura Y (
2000
) Bilateral receptive field neurons and callosal connections in the somatosensory cortex.
Philos Trans R Soc Lond B Biol Sci

355
:
267
–273.
Koralek K, Olvarria J, Killackey H (
1990
) Areal and laminar organization of corticocortical projections in the rat somatosensory cortex.
J Comp Neurol

299
:
133
–150.
Krupa D, Ghazanfar A, Nicolelis M (
1999
) Immediate thalamic sensory plasticity depends on corticothalamic feedback.

96
:
8200
–8205.
Krupa DJ, Matell MS, Brisben AJ, Oliveira LM, Nicolelis MAL (
2001
) Behavioral properties of the trigeminal somatosensory system in rats performing whisker-dependent tactile discriminations.
J Neurosci

21
,
5752
–5763.
Lewine JD, Doty RW, Astur RS, Provencal SL (
1994
) Role of the forebrain commissures in bihemispheric mnemonic integration in macaques.
J Neurosci

14
:
2515
–2530.
Martin JH (
1991
) Autoradiographic estimation of the extent of reversible inactivation produced by microinjection of lidocaine and muscimol in the rat.
Neurosci Lett

127
:
160
–164.
Mello NK (
1965
) Mirror-image reversal in pigeons.
Science

149
:
1519
–1520.
Milani H, Steiner H, Huston J (
1989
) Analysis of recovery from behavioral asymmetries induced by unilateral removal of vibrissae in the rat.
Behav Neurosci

103
:
1067
–1074.
Noble J (
1968
) Paradoxical interocular transfer of mirror-image discriminations in the optic chiasm sectioned monkey.
Brain Res

10
:
127
–151.
Olavarria J, van Sluyters R, Killackey H (
1984
) Evidence for the complementary organization of callosal and thalamic connections within rat somatosensory cortex.
Brain Res

291
:
364
–368.
Pidoux B, Verley R (
1979
) Projections on the cortical somatic I barrel subfield from ipsilateral vibrissae in adult rodents.
Electroencephalogr Clin Neurophysiol

46
:
715
–726.
Shuler MG, Krupa DJ, Nicolelis MAL (
2001
) Bilateral integration of whisker information in the primary somatosensory cortex of rats.
J Neurosci

21
:
5251
–5261.
Sperry W (
1964
) The great cerebral commissure.
Sci Am

210
:
42
–52.
Vincent S (
1912
) The function of the vibrissae in the behavior of the white rat.
Behav Monogr

1
:
7
–81.
Welker WI (
1964
) Analysis of sniffing of the albino rat.
Behaviour

12
:
223
–244.
Welker E, Hoogland P, Van der Loos H (
1988
) Organization of feedback and feedforward projections of the barrel cortex: a PHA-L study in the mouse.
Exp Brain Res

73
:
411
–435.
White E, DeAmicis R (
1977
) Afferent and efferent projections of the region in mouse SmI cortex which contains the posteromedial barrel subfield.
J Comp Neurol

175
:
455
–482.