Absence of plasticity of the frequency map in dorsal cochlear nucleus of adult cats after unilateral...

Absence of Plasticity of the FrequencyMap in Dorsal Cochlear Nucleusof Adult Cats After Unilateral

Partial Cochlear Lesions

R. RAJAN1* AND D.R.F. IRVINE2

1Department of Physiology, Monash University, Clayton, Victoria 3169, Australia2Department of Psychology, Monash University, Clayton, Victoria 3168, Australia

ABSTRACTIn adult animals, lesions to parts of the auditory receptor organ, the cochlea, can produce

plasticity of the topographic (cochleotopic) frequency map in primary auditory cortex and arestricted or patchy plasticity in the auditory midbrain. This effect is similar to the plasticityof topographic maps of the sensory surface seen in visual and somatosensory cortices afterrestricted damage to the appropriate receptor surface in these sensory systems. There isdispute about the extent to which subcortical effects contribute to cortical plasticity. Here, wehave examined whether topographic map plasticity similar to that seen in the auditory cortexand the midbrain is observed in the adult auditory brainstem. When partial cochlear lesionswere produced in the same manner as those that were produced in the cortex and midbrainstudies, we found no plasticity of the frequency map in the dorsal cochlear nucleus (DCN).Small regions of the DCN that were deprived of their normal, most sensitive frequency(characteristic frequency; CF) input by the cochlear lesion appeared to have acquired new CFsat frequencies at or near the edge of the cochlear lesion. However, examination of thresholds atthe new CFs established that the changes simply reflected the residue of prelesion input tothose sites: The patterns of CF thresholds were very well predicted by simple calculations ofthe patterns that were expected from such residual input. The results of this study suggestthat the DCN does not exhibit the type of plasticity that has been found in the auditory cortexand midbrain; therefore, it does not account for the changes in responsiveness observed in thehigher level structures under similar experimental conditions. J. Comp. Neurol. 399:35–46,1998. r 1998 Wiley-Liss, Inc.

Indexing terms: tonotopic maps; subcortical plasticity; cochlea; cortical plasticity; adults

When the neural outflow from restricted parts of sensoryepithelia in adult animals is eliminated by damage eitherto neurons or to receptors, neurons in appropriate regionsof the topographic cortical representations (maps) of thereceptor surface dynamically acquire sensitivity to recep-tor surface regions adjacent to the damage. Such plasticity,resulting in an expanded cortical representation of theadjacent receptor surface regions, occurs in visual (Kaas etal., 1990; Heinen and Skavenski, 1991; Chino et al., 1992;Gilbert and Weisel, 1992; Darian-Smith and Gilbert, 1995;Schmidt et al., 1996), somatosensory (see, e.g., Merzenichet al., 1983, 1984; Rasmusson, 1982), and auditory (Robert-son and Irvine, 1989; Rajan et al., 1993; Schwaber et al.,1993) systems. In the latter, restricted lesions of thetonotopically organized cochlea can produce plasticity ofthe cochleotopic frequency map in primary auditory cortex(AI). AI neurons that are deprived of their normal, most

sensitive frequency (characteristic frequency; CF) inputacquire sensitivity to frequencies represented at cochlearsites adjacent to the lesion (Robertson and Irvine, 1989;Rajan et al., 1993; Schwaber et al., 1993). Critically,consideration of thresholds at the new CFs in the reorga-nized map established that the effects reflected plasticchanges rather than simply being the residue of prelesioninput (Rajan et al., 1993). Furthermore, the fact that newCFs following reorganization of the cortical map after a

Grant sponsor: National Health and Medical Research Council of Austra-lia; Grant number: 920483.

*Correspondence to: Dr. R. Rajan, Department of Psychology, MonashUniversity, Clayton, Victoria 3168, Australia.E-mail: ramesh.rajan@med.monash.edu.au

Received 22 January 1998; Revised 7 May 1998; Accepted 11 May 1998

THE JOURNAL OF COMPARATIVE NEUROLOGY 399:35–46 (1998)

r 1998 WILEY-LISS, INC.

long survival period can differ from those seen immedi-ately after the lesion indicates that the plastic changes incortex do not involve merely reduced thresholds at postle-sion CFs (Irvine and Robertson, 1990).

There is dispute about the extent to which corticalplasticity reflects subcortical effects. In the visual system,either a limited plasticity has been found at the thalamus(Eysel et al., 1980), or little change was seen in thethalamic lateral geniculate nucleus (LGN) in associationwith substantial cortical plasticity (Gilbert and Weisel,1992; Darian-Smith and Gilbert, 1995). In the somatosen-sory system, subcortical plasticity has been suggested(Pons et al., 1991) or observed in the spinal cord (Basbaumand Wall, 1976; Devor and Wall, 1981), brainstem (Millaret al., 1976), or thalamus (Garraghty and Kaas, 1991;Rasmusson, 1996). However, other studies have failed tofind plasticity in the spinal cord (see, e.g., Pubols andGoldberger, 1980; Pubols and Brenowitz, 1982).

In the auditory system, we found a ‘‘patchy’’ plasticity inthe midbrain nucleus, the inferior colliculus (IC), such thatsome parts of the central nucleus of the IC expressedplasticity, whereas others did not (Irvine and Rajan, 1994).The only other study on adult subcortical auditory plastic-ity reported that post-cochlear-lesion changes in the fre-quency map in the brainstem dorsal cochlear nucleus(DCN) could be explained as residue of prelesion input anddid not account for cortical plasticity (Kaltenbach et al.,1992, 1996). This conclusion must be qualified for tworeasons. First, in the study by Kaltenbach et al. (1992),cochlear damage was produced with loud sound, which hasnot yet been demonstrated to induce plasticity of fre-quency maps anywhere in the auditory pathway. Second,auditory cortical plasticity appears to be expressed onlywhen certain boundary conditions are met (cf. Rajan andIrvine, 1996). Principal among these conditions is that, inthe case of damage to restricted parts of the cochlea, someparts of the cochlea have suffered severe damage to, ordegeneration of, the receptor epithelium and cochlearafferent fibers sufficient to deprive cortical regions of all ormost of their prelesion input (Rajan and Irvine, 1996). Inthe DCN study, in which cochlear damage was quantifiedsolely by histology of receptors, only one case (see Fig. 2 inKaltenbach et al., 1992) showed severe histological dam-age sufficient to suggest that the boundary conditions for(cortical) plasticity had been met. Furthermore, in this onecase, significant histological damage to particular receptorcells was seen in apical cochlear regions, yet there ap-peared to be no effect on the DCN map for the lowfrequencies represented in the cochlear apex. The ad-equacy of the histological index must also be questionedwhen it is noted that significant changes in a DCN mapwere observed with minimal histological damage to recep-tors (see Fig. 3 in Kaltenbach et al., 1992).

To resolve these issues, we examined whether plasticityof DCN maps occurred after cochlear lesions that werevery similar to lesions that produced cortical plasticity. Thiswould also help determine whether the plasticity found in theIC reflected, at least in part, plasticity expressed in theDCN, which provides a separate and distinct pathway ofthe multiple brainstem inputs to the IC.

MATERIALS AND METHODS

Experiments were carried out on adult domestic cats byusing the procedures approved by the Monash University

Standing Committee on Ethics in Animal Experimenta-tion and by conforming to guidelines set out in theNational Institutes of Health Guide for the Care and Useof Laboratory Animals and the policy of the Society forNeuroscience. Animals were allocated to two groups: onegroup of three normal animals in which no procedureswere carried out prior to the terminal experiment to recordfrom DCN and one group of four test animals in whichmechanical lesions were made to the left cochlea somemonths prior to DCN recording.

All experimentation was carried out in a shielded,sound-attenuating room. Anesthesia was induced by anintraperitoneal injection of sodium pentobarbitone (Nem-butal; 40 mg/kg body weight), and the animal was given asingle intramuscular injection of atropine sulfate (0.2 ml of600 µg/ml solution) to reduce mucous secretion. Theanimal was wrapped in a thermostatically controlledheating blanket, and rectal temperature was monitoredcontinuously. In the lesioning phase (test animals only) ofthis study, only the single dose of Nembutal was given:Surgery was completed within 1 hour, and the animal wasallowed to recover as detailed elsewhere (Rajan et al.,1993).

Procedures for cochlear lesions and recovery in testanimals were carried out as described previously (Rajan etal., 1993). In brief, the left round window (RW) wasexposed surgically, and a glass micropipette was advancedthrough the RW and basilar membrane a number of timesto create a loss in hearing sensitivity. This loss wasassessed by measuring visual detection thresholds (VDTs)for the compound action potential (CAP) audiogram atfrequencies from 0.5 kHz to 40 kHz (Rajan et al., 1991)prior to and after the lesion. Animals recovered for 2months (CML 9420 and CML 95105), 5 months (CML9418), or 5.5 months (CML 9466).

In the terminal DCN recording phase in all animals,after induction of a surgical level of anesthesia, a radialvein was cannulated to administer a single dose of Decad-resson (sodium dexamethasone phosphate) and for subse-quent administration of Nembutal at approximately 2–3mg/kg/hour to maintain deep anesthesia. Depth of anesthe-sia was monitored through continuous recording of therectal temperature, the electrocardiogram, and the electro-myelogram activity from forearm muscles and by regularchecks of pupillary dilatation and the pinch-withdrawalreflex. Body temperature was maintained at 37.5 6 0.5°Cby a thermostatically controlled warming blanket wrappedaround the animal and was regulated by feedback througha rectal temperature probe. Surgery was carried out toimplant stainless-steel electrodes against the RW mem-brane of both cochleas to measure the CAP audiogrambilaterally in all control animals and in three of the fourtest animals (Rajan et al., 1991). In one test animal, theaudiogram was measured only from the lesioned ear.

A posterior craniotomy was made (Rajan, 1997) toexpose the DCN. In all three control animals and in twotest animals (CML 9466 and CML 95105), the craniotomywas made unilaterally to eventually expose the left cochlearnucleus (CN). In the other two test animals (CML 9418and CML 9420), the craniotomy was made large enough toexpose both CN. The cerebellum overlying the CN wasremoved by suction. To record from the DCN, a glass-insulated tungsten microelectrode (2–5 MV at 1 kHz) wasaligned 37° off the midsagittal plane and 37° off the coronalplane (see Fig. 1 in Spirou et al., 1993) to penetrate the

36 R. RAJAN AND D.R.F. IRVINE

DCN near its dorsomedial tip and to run along the longstrial axis of the nucleus. With this approach, the electrodetraverses the tonotopic axis of the DCN (Spirou et al.,1993; Fig. 2A in the present study). The electrode wasadvanced rapidly in a ventral direction into the depth ofthe DCN, with only random sampling of activity duringthis forward penetration, until the neuronal clusters thatwere most sensitive to a low frequency (,4–8 kHz) werefirst encountered. The electrode was left in this position for<10 minutes to allow any compression of tissue to settle.Then, the most sensitive frequency (characteristic fre-quency; CF) and the threshold at CF of the neuronalcluster (‘‘cluster’’ recording) were determined audiovisu-ally. Finally, quantitative data to tone bursts over a rangeof frequencies and intensities were obtained to define theboundaries of the excitatory response area (ERA) andresponse strength within the ERA, as detailed previously(Rajan et al., 1993; Rajan, 1997). The electrode was thenmoved back out of the DCN in steps of 50 µm, andrecordings from neuronal clusters were made every 100–200 µm. When possible, data were obtained from a numberof such penetrations.

Stimuli were pure tone bursts (50 msec duration, 4 msecrise-fall times) that were presented at 2/second to the earipsilateral to the DCN from which recordings were made.Methods of stimulus generation, delivery, and calibrationwere as described previously (Rajan et al., 1991, 1993). Intwo test animals, data were obtained from the DCNipsilateral and that contralateral to the lesioned cochlea.In the other two test animals, data were obtained onlyfrom the DCN ipsilateral to the lesioned cochlea. Forconvenience, in all lesioned animals, the DCN ipsilateralto the lesioned cochlea will be referred to as ipsilesionaland the DCN contralateral to the lesioned cochlea will bereferred to as contralesional.

The experiment was terminated with an overdose ofNembutal, which was administered through the i.v. can-nula. In some experiments, one or both cochleas from thetest animals were then removed for histology (Rajan et al.,1993). In some experiments from control and test animals,

prior to the overdose, the tungsten microelectrode wasreadvanced into the DCN, and electrolytic lesions weremade at selected points along the length of a successfultrack. In these cases, the DCN was also recovered postmor-tem for histology.

RESULTS

Effect of lesions on cochlear hearingsensitivity in test animals

CAP audiograms in the lesioned left cochleas of the testanimals, which are illustrated in Figure 1A, show that allfour animals had CAP threshold losses commencing atsome intermediate frequency and increasing to higherfrequencies. Generally, the threshold losses consistentlycommenced from a frequency of about 12–15 kHz, withlarge losses ($ 20 dB relative to normal thresholds; one ofthe boundary conditions we have shown to be necessary forcortical plasticity; Rajan and Irvine, 1996) commencingfrom about 15–19 kHz and with losses increasing at higherfrequencies. In two animals (CML 9420 and CML 9466), noCAP could be recorded at frequencies $ 24 kHz. In theother two animals (CML 9418 and CML 95105), the CAPwas absent at frequencies of $ 30–32 kHz. This pattern ofCAP thresholds is very similar to those recorded in ourcortex study (Rajan et al., 1993). In contrast, thresholds incontrol animals and in the unlesioned cochleas of the testanimals consistently lay within the normative range. Thisis illustrated in Figure 1B for the unlesioned cochleas ofthree test animals. Histological examination of the cochleasshowed effects similar to those described previously withsuch lesions (Rajan et al., 1993; Rajan and Irvine, 1996): Aquite sharp edge extending over a few hundred micronsseparated the region of normal cochlea (as determined bylight microscopy) located apically from a basal region inwhich the organ of Corti had collapsed completely, withdestruction of all hair cells and with only a few spiralganglion cells surviving (, 5–10% of the normal comple-ment). In the intermediate region, there was graded

Fig. 1. Cochlear sensitivity in lesioned and unlesioned cochleas oftest animals at the time of dorsal cochlear nucleus (DCN) mapping,with respect to normal cochlear sensitivity. Data shown are the visualdetection thresholds (VDTs) for the compound action potential (CAP)of the auditory nerve to tone bursts. The hatched region represents therange 6 1.64 standard deviations from the mean of a large database ofnormative CAP VDTs from the laboratory (cf. Rajan et al., 1991), and

the dotted line represents the 20 dB . mean normative CAP VDT.A: Data from the lesioned left (LT) cochleas of all four test animals(CML 9418, 9420, 9466, and 95105). B: Data from the unlesioned right(RT) cochleas of the three test animals from which these data wereobtained (CML 9418, 9420, and 95105). Each symbol represents datafrom one animal, identified at the top left of the plot. NR, no response.

ABSENCE OF PLASTICITY IN ADULT DORSAL COCHLEAR NUCLEUS 37

damage to the hair cells. Given the similarity of thesehistological effects to those described previously (Rajan etal., 1993; Rajan and Irvine, 1996), these data are notpresented here.

Tonotopic order in DCN of controland test animals

In control animals, as described by Spirou et al. (1993),neuronal clusters in ventral DCN had low CFs and clus-ters in dorsal DCN had high CFs, with a generallymonotonic increase in CF in between. This is illustrated inFigure 2A for the four penetrations in control animal NCN9457. Although all four penetrations did not record theentire CF range represented in DCN (Spirou et al., 1993), atonotopic gradient of CFs from 8 kHz to at least 32 kHzwas always obtained. The differences in the depth at whichany particular CF was obtained across the four penetra-tions reflects the change in shape and curvature of theDCN and, thus, the exact starting point of the penetrationalong the tonotopic axis of DCN (Spirou et al., 1993; Fig. 3of present study). In three penetrations, after CFs of 32–40kHz had been obtained dorsally in DCN, more dorsallylocated neuronal clusters were broadly tuned for frequency(‘‘A’’ points) or unresponsive (‘‘X’’ points). In the fourth penetra-tion, the most dorsal recording site yielded a CF of 32 kHz.

The tonotopic sequence in the ipsilesional left DCN inone test animal (CML 9420) is illustrated in Figure 2B.Again, low CFs were obtained in ventral DCN, with a CF of, 5 kHz obtained in at least two of the five penetrations. Atonotopic CF sequence was obtained in each penetration asthe electrode was moved dorsally, until a CF of about12–18 kHz was first encountered. Thereafter, more dor-sally, the CF-distance plot flattened out in each penetra-tion as CFs of about 10–16 kHz were obtained for somedistance, the extent of which varied between differentpenetrations but was between 0.5 mm and 1 mm. Moredorsal to this area, neuronal clusters did not respond toauditory stimuli at all (X points). No CF greater than

about 18 kHz was recorded in any of the five penetrationsin this animal.

The CF gradient in the ipsilesional DCN in all four testanimals is shown in Figure 3E–H and can be comparedwith the gradient seen for two control animals (Fig. 3A,B)and for the contralesional DCN of two lesioned animals(Fig. 3C,D). In each animal, the different penetrationswere aligned with respect to one another at the distance atwhich a CF of 10 kHz was first encountered. By comparingFigure 3A with Figure 2A, it can be seen that, when thisalignment procedure is carried out, the tonotopic gradientsin different penetrations in the one animal overlie verywell with respect to the CF recorded along the strial axis ofDCN. In the two normal cases and the two contralesionalcases illustrated in Figure 3A–D, there is a fairly mono-tonic progression of CF along the strial axis of DCN.Although the full CF sequence in DCN (Spirou et al., 1993)was not recorded in all penetrations in all four cases, CFsup to about 32 kHz were always obtained.

In contrast to this normal tonotopic CF gradient, the CFgradient in ipsilesional DCN of the four test animals (Fig.3E–H) was clearly abnormal in DCN dorsal to the 10 kHzCF point. Low CFs were always found in ventral DCN, andthere was a normal CF progression as the electrode wasmoved more dorsally. In two animals (CML 9420 and CML95105; Fig. 3E,F), this progression was maintained until aCF of about 16 kHz was first encountered. Thereafter, theCF-distance plot in these two animals flattened out, indi-cating that a CF of about 16 kHz was obtained over somedistance, or it dipped down to even lower CFs. This effectoccurred over a distance of , 1 mm that varied betweenthe two animals and between different penetrations in theone animal, then it was followed by more dorsal DCNregions, where neurons were either broadly tuned forfrequency (A points) or, most often, unresponsive to sound(X points). In CML 9418 (Fig. 3G), the CF-distance mapflattened out at about 20–22 kHz for some distance in twopenetrations before unresponsive points were recorded

Fig. 2. Progression of characteristic frequency (CF) with distancein the left dorsal cochlear nucleus (DCN) in control animal NCN 9457(A) and in test animal CML 9420 (B), in which a lesion was made inthe left cochlea 2 months prior to the DCN mapping. The hatchedregion labeled A indicates neuronal clusters that were responsive totone-burst stimulation but very weakly, such that a clear CF could notbe assigned. The shaded region labeled X indicates neuronal clustersthat were unresponsive to tone-burst stimulation. Each symbol indi-cates data from a single penetration along the strial axis of DCN.

Symbols surrounded by rectangles indicate that neuronal clusters atthat point were broadly tuned for frequency even near threshold; insuch cases, the symbol has been placed at the frequency that elicitedthe strongest response at threshold, whereas the box indicates therange of frequencies over which the cluster responded within 5 dB ofthreshold. Each plot is to be read from right to left to see the CFprogression from low CFs ventral in the penetration and high CFsdorsally.

38 R. RAJAN AND D.R.F. IRVINE

more dorsally. In the third penetration (Fig. 3G, opencircles), the flattening out of the CF-distance plot wasmixed with dips to much lower CFs in the plot, particularlyat the most dorsal locations. Finally, in CML 9466 (Fig.3H), the CF-distance plot for each of two penetrationsflattened out at about 22 kHz for some distance beforeunresponsive points were recorded more dorsally.

The differences between the normal CF sequences (fromboth control left DCN and from the contralesional DCN oftest animals) and sequences in the ipsilesional DCN of testanimals can be appreciated more easily by comparing theoverall CF gradients in these different cases. This wasdone by fitting a line of best approximation to all data ineach of the two subcategories of ipsilesional DCN data (thetwo ipsilesional DCN cases with CF-distance plot flatten-ing at about 16 kHz and the two cases with CF-distanceplot flattening at about 22 kHz, as detailed above) and thetonotopic sequence in the total normal DCN set (fourcontrol DCN and two contralesional DCN). To fit the line ineach of these categories, CF data from all penetrations,aligned to the 10-kHz CF point, were overlaid and sortedaccording to distance. The CF-distance relationship ineach category was then approximated by a Beiter-splinefunction, using a low spline parameter (kept constantacross all three categories) to keep smoothing to a mini-mum. The qualitative Beiter-spline function was preferredto a more quantitative function, such as a polynomial,because, although polynomials were good descriptors whenCF-distance plots were monotonic, they were poor descrip-tors when there were dips in the CF-distance relationship,as in CML 9418 and, to a lesser extent, CML 9420. Infitting this spline function, A and X points in all three datasets were not used, i.e., only points with clearly assignableCFs were used to determine the CF-distance relationship.To illustrate the fact that the Beiter-splines were a goodapproximation of the overall aligned CF gradients for eachgroup, Figure 3I–K shows the spline functions fitted toeach of the three groups.

Finally, to compare the tonotopic sequences across thethree groups, the spline fits alone are shown in Figure 3L.Commencing ventrally in DCN, the tonotopic sequence inthe two subcategories of ipsilesional DCN in test animalsinitially corresponds closely to that of DCNs receivinginput from unlesioned ears (the ‘‘normal’’ DCNs, consistingof the control, unlesioned animals, and the contralesionalDCNs in the test animals). However, at distances about0.5–1.0 mm more dorsal to the 10-kHz CF point, theCF-distance plots of the two subcategories of ipsilesionalDCNs diverge from the normal tonotopic sequence as theCF-distance plots in the former cases flatten out at about16 kHz, or about 22 kHz. In the case in which theCF-distance plot flattens out at about 16 kHz, this effect isseen for only a relatively short distance of up to about 1.25mm dorsal to the 10-kHz CF point (a distance correspond-ing in the normal CF-distance plot to a normal CF of about30 kHz) before cells that are unresponsive to auditorystimuli are encountered more dorsally. In the case in whichthe curve flattens out at about 22 kHz, this is followed bythe CF-distance plot dipping down to much lower CFs (duespecifically to one penetration in one animal, CML 9418).Thus, in this subcategory (and, specifically, in CML 9418),CFs appear to be recorded over the same extent of DCN asin normal DCN, although there is a marked difference inthe tonotopy in this subcategory compared with the nor-

mal DCN tonotopy, which extends to 40 kHz (the upperfrequency limit in our testing system).

In summary, the tonotopy in ipsilesional DCN in testanimals shows features that are somewhat similar to someeffects in contralateralAI (Rajan et al., 1993) after mechani-cal cochlear lesions of the same type used here, producingdestruction or desensitization of high-frequency regions ofthe cochlea. In ipsilesional DCN regions in which high CFs(corresponding to frequencies represented in the damagedportion of the cochlea) would be found normally, there is (inmost cases) an ‘‘expanded’’ representation of lower frequen-cies, generally of frequencies close to or at the edge of thecochlear lesion.

Thresholds at CF in DCN in controland lesioned animals

To establish whether the postlesion changes in the CFmap are a manifestation of plasticity or simply reflect theresidue of prelesion input (Rajan et al., 1993; Rajan andIrvine, 1998), it is necessary to consider thresholds at the‘‘new’’ CFs. In the residue argument, the flattening of theCF-distance plot reflects the fact that, normally, there isconvergence and divergence between different frequencychannels in central auditory pathways. In CF regions inwhich the normal CF input has been destroyed or desensi-tized by the cochlear lesion, the ‘‘new’’ CF could simply bethe most sensitive input remaining after elimination of theformer CF input. The fact that ‘‘new’’ CFs were always atfrequencies lower than those expected in the DCN regionscorresponding to the lesion reflects the facts that thelow-frequency slopes of ERAs are shallower than high-frequency slopes and that our cochlear lesions producedlosses that commenced at some intermediate frequencyand increased with increasing frequency. This would re-sult in the observed ‘‘expanded’’ representation of a lowfrequency near the edge of the cochlear lesion.

It is to be noted that, according to the residue argument,the residual responses (due to inputs from frequenciesother than the normal CF inputs), by definition, would notbe the prelesion inputs to which the neuron was mostsensitive. Thus, thresholds at these postlesion residualCFs should be higher than those expected of neurons forwhich this was the normal CF. Furthermore, if all high-CFneurons lost their CF input and were left with a similarresidual input from a cochlear lesion-edge frequency, thenit would be possible, from the ERAs (see Materials andMethods) in normal neurons, to predict the progression ofCF thresholds expected to be encountered across the DCNregions expressing the ‘‘new’’ residual CF. In our AI, studywe have shown how this procedure could be applied (Rajanet al., 1993). Here, we apply the same procedure to thepresent DCN data.

Threshold data at CF from all four ipsilesional DCNs areshown in Figure 4A–D, in which they are plotted withpenetrations aligned at the 10-kHz CF point, like theCF-distance data in Figure 3.

In all penetrations in all four ipsilesional DCNs, CFthresholds remained low (# 20 dB) only up to and for shortdistances beyond the 10-kHz CF point. From distances of,0.5 mm more dorsal to this point progressing to evenmore dorsal DCN, CF thresholds rise steeply and, for themost part, monotonically. These effects can be contrastedwith CF-threshold data from the DCN in two controlanimals (Fig. 4E,F) and from the contralesional DCN intwo test animals (Fig. 4G,H; note that Fig. 4E–H illus-

ABSENCE OF PLASTICITY IN ADULT DORSAL COCHLEAR NUCLEUS 39

Figure 3

trates the same cases shown in Fig. 3 for CF-distanceplots). In all four plots, CF thresholds remain low (# 20dB) for distances up to at least 1.0 mm more dorsal to the10-kHz CF point. Thereafter, although CF thresholdsincrease, the rate of increase in normal DCN (controlDCNs and contralesional DCNs in two test animals)generally appears to be much shallower than that in theipsilesional DCN of test animals.

Direct comparison of the threshold-distance gradient inthe ipsilesional DCN against that in the normal pool wasattempted by using the same analysis that was applied tothe CF-distance data in Figure 3.

A line of best approximation was fitted to all data in eachof the two subcategories of ipsilesional DCN data (the twoipsilesional DCN cases with CF-distance plot flatteningabout 16 kHz and the two cases with CF-distance plotflattening about 22 kHz) and to the data in the totalnormal DCN set (control and contralesional DCNs). TheBeiter-spline functions fitted to each of the three groupsare shown in Figure 4I–K to illustrate the fact that thespline functions were a good approximation of the overall,aligned, CF-threshold gradients for each group. Figure 4Lcompares the three spline fits. Ventral to the 10-kHz CFpoint, the spline fits to the threshold-distance data in allthree categories of DCN data overlie quite well. Thiscongruity holds true with dorsal progression in DCN untila position almost 0.5 mm more dorsal to the 10-kHz CFpoint. Thereafter, the lines fitted to the two subcategoriesof ipsilesional DCN data diverge markedly from the linefitted to the total normal DCN pool, as thresholds inipsilesional DCN increase much more steeply than innormal DCN. Comparison of Figure 4L and Figure 3Lshows that, in the region of the ipsilesional DCN in whichthe CF-distance plot diverges from the normal CF-distanceplot, CF thresholds also increase steeply.

Predictions of CF thresholdsby the residue hypothesis

The marked divergence of CF thresholds in the dorsalregion where the CF-distance plots flattened out in ipsile-sional DCN is very different from the effects seen in AIplasticity, in which CF thresholds remain low in the regionwhere the CF-distance plots flatten out (Robertson andIrvine, 1989; Rajan et al., 1993; Rajan and Irvine, 1998).This suggests that the DCN effects are not similar to AIplasticity and could simply reflect the residue of prelesioninputs (Rajan et al., 1993; Rajan and Irvine, 1998).

The argument that the CF-distance plots in the ipsile-sional DCN in test animals are explicable as the residue ofprelesion input is strengthened by comparison of CFthresholds measured in the DCN in these cases with theCF thresholds predicted from the residue argument. Thelatter thresholds were derived from the ERAs measured inthe normal pool (i.e., control DCNs and contralesionalDCN in two test animals) in the manner explained below.Figure 5A–D illustrates the CF-threshold predictions fortwo control DCNs and two contralesional DCNs in testanimals for the case in which the residue of prelesion inputresulted in an expanded representation of a frequency of16 kHz. In these plots, for all CFs up to 16 kHz, thethresholds plotted are those measured at CF (i.e., for CFsup to 16 kHz, the data are exactly those shown in theappropriate plots in Fig. 4). Thereafter, for points withCFs . 16 kHz, the threshold plotted is not that at CF butrather threshold at 16 kHz, which was obtained from thequantitative measurements of the ERA for higher-CFclusters. Similar calculations were also made in the othertwo control animals (not illustrated).

Similarly, Figure 5E–H illustrates the CF thresholdspredicted by the residue argument in the case in whichthere is an apparent ‘‘expanded’’ representation of 22 kHz.In these plots, which were also derived from normal data,thresholds plotted for all CFs up to 22 kHz are those at CF(i.e., for CFs up to 22 kHz, data plotted are exactly those inappropriate plots in Fig. 4). Thereafter, for points with CFs. 22 kHz, the threshold plotted is that at 22 kHz, which,again, was obtained from the ERA measured in the higher-CF clusters.

In both rows, it can be seen that the residue argumentwould predict that, beyond the ‘‘lesion-edge’’ CF (i.e., 16kHz for Fig. 5A–D and 22 kHz for Fig. 5E–H), there wouldbe a sharp increase in CF thresholds as recordings weremade from neuronal clusters located more dorsally inDCN. These predicted effects can be compared with theactual progression of CF thresholds recorded in the leftDCN of the lesioned animals. (In making this comparison,it should be noted that, in the two subsets of test DCNdata, in the region where the CF-distance plots ‘‘flattenedout,’’ the CF was not always exactly at 16 kHz or exactly at22 kHz. Hence, the progression of CF thresholds calcu-lated here for cases in which the progression was of CFs at16 kHz or at 22 kHz should be treated only as approxima-tions of the progression expected in the test DCN subsets.)Like in Figures 3 and 4, this was done by fitting aBeiter-spline function to provide a single approximation ofall data in a particular pool. Here, all data for thepredictions for a residual 16 kHz ‘‘expanded’’ representa-tion were pooled, and a Beiter-spline function was fitted tothose data. A similar procedure was carried out on data forthe predictions for a residual 22 kHz ‘‘expanded’’ representa-

Fig. 3. Progression of characteristic frequency (CF) with distancein normal and ipsilesional dorsal cochlear nucleus (DCN). A–D: Normal(i.e., left) DCN, in two control animals (A,B) and contralesional (i.e.,right) DCN in two test animals (C,D). E–H: DCN ipsilateral to thelesioned cochlea in test animals. The animal is identified in the topright corner of each plot from A to H. The layout of the ordinate is thesame as that described for Figure 2. For the abscissa, in each animal,the different penetrations were aligned with respect to one another atthe distance at which a CF of 10 kHz was first encountered. Theabscissa plots distance in the DCN relative to this point, with positivedepths indicating points farther ventral along the DCN strial axisthan the 10-kHz CF point and negative values that indicate moredorsally located points. Otherwise, the layout here is the same as thatfor Figure 2, with each symbol indicating data from a single penetra-tion along the strial axis of DCN. I–L: Comparison of the CF-distanceprogression in normal DCN vs. that in each subcategory of ipsilesionalDCN. In I–L, the full line presents the Beiter-spline of best approxima-tion for the trend of data (symbols) in that plot. I presents the data forthe ipsilesional DCN for the two test animals in which the CF-distanceplot flattened out at ,16 kHz (CML 9420, shown in E; CML 95105,shown in F). J presents the data for the ipsilesional DCN for the twotest animals in which the CF-distance plot flattened out at ,22 kHz(CML 9418, shown in G; CML 9466, shown in H). K presents data fromthe four normal DCNs shown in A–D and from another two controlanimals. Data from the different animals for each plot were aligned atthe distance of the 10-kHz CF point and then sorted according todistance. Then, the Beiter-spline approximation was fitted by using alow spline parameter that was kept constant across all three plots(I–K). These fitted lines alone are compared in L, in which CMLs 16indicates the line fitted in I, CMLs 22 indicates the line fitted in J, andNORMALS indicates the line fitted in K. Note the different abscissascale in L compared with the other plots.

ABSENCE OF PLASTICITY IN ADULT DORSAL COCHLEAR NUCLEUS 41

Figure 4

tion. Figure 5I compares the spline function fitted in thecase of the residual 16-kHz ‘‘expanded’’ representationwith the spline functions fitted to the CF-threshold progres-sion seen in the normal data pool (the same spline fit as inFig. 4K) and in ipsilesional DCN in the two test animals inwhich the CF-distance plot flattened out at about 16 kHz(the same spline fit as in Fig. 4I). The sequence of 16-kHzCF thresholds predicted by the residue argument matchesquite closely the way in which the sequence of CF thresh-olds actually recorded in the ipsilesional DCN in the twotest cases diverged from the sequence in the normal pool.The only marked difference between the predicted functionand that observed in the lesion animals is very dorsally inthe DCN, where the ipsilesional DCN becomes unrespon-sive. This must reflect test DCN regions that are com-pletely deprived of any input (where the residue argumentstill predicts some low-CF input).

Similarly, Figure 5J compares the Beiter-spline fit to theCF-threshold predictions of the residual 22-kHz CFs argu-ment against the spline fit to the CF-threshold progressionseen in the normal data pool (again, the same spline fitused in Fig. 4K) and in the ipsilesional DCN in the two testanimals in which the CF-distance plot flattened out atabout 22 kHz (the same spline fit used in Fig. 4J). There isless congruence between the predicted function and thelesion data here than in Figure 5I. However, it should benoted that the CF thresholds that were actually recordedin the ipsilesional cases shown here, in fact, were worsethan those predicted by the residue argument. In part, thiscould reflect the fact that the CF sequence in the ipsile-sional DCN in these two test animals (see Fig. 3J) actuallyshows irregular dips to lower CFs, which would havehigher ‘‘residual’’ thresholds than a frequency of 22 kHz ina formerly high-CF neuronal cluster. It could also reflectthe fact that the observed CF thresholds in the ipsilesionalDCN in these two animals appear to be slightly offset fromthe normative data, even for lower CFs that are just . 10kHz (Fig. 5J). If this small offset of about 5–10 dB is

corrected for, then there is quite good correspondencebetween the 22-kHz CF thresholds predicted by the resi-due argument and those that were actually recorded inthese ipsilesional DCNs in which the CF-distance plotflattens out at about 22 kHz. In any case, the fact that CFthresholds in these two animals are worse than thosepredicted by the residue argument would at least arguestrongly against the idea that the CF gradients observedin these animals reflect plasticity.

DISCUSSION

Following unilateral, partial cochlear lesions in adultanimals, the tonotopic map in the ipsilesional DCN isaltered. In at least some penetrations in all lesion animals,neurons in ipsilesional DCN regions where high CFswould normally be found have a new CF, generally at afrequency that is close to or at the edge of the cochlearlesion (lesion-edge frequencies). This results either in aflattening out of the CF-distance plot or a flattening outmixed with irregular dips to much lower frequencies in theCF-distance plot. Thus, in the ‘‘deprived’’ DCN, there is an‘‘expanded’’ representation of lower frequencies close to theedge of the cochlear lesion, with varying degrees of scatterwith regard to the frequency range showing this expandedrepresentation. The distance of deprived DCN over whichthis change occurs varies between penetrations even in theone animal. This effect is similar to that described byKaltenbach et al. (1992) after acoustically induced cochlearlesions.

The flattening out effect is reminiscent of the flatteningout of the CF-distance plot (and expanded representationof lesion-edge frequencies) seen in contralesional AI aftersimilar cochlear lesions (Rajan et al., 1993). However,there are two major respects in which the DCN effectdiffers from the cortical effect. The first difference is that,over the region in which the CF-distance plot in ipsile-sional DCN varies from the normal CF-distance progres-sion, there is a systematic and generally-monotonic in-crease in thresholds at the ‘‘new’’ CFs. This is very differentfrom the effect seen in contralesional AI, in which thresh-olds remain low when progressing from normal AI regionsinto the AI region containing the new CFs at lesion-edgefrequencies (Rajan et al., 1993; Rajan and Irvine, 1998)and are not significantly different from normal thresholdsfor those CFs. This difference in the progression of CFthresholds in ipsilesional DCN and contralesional AI sug-gests that the DCN effects do not represent plasticity of thetonotopic organization, as in contralesional AI, but simplythe residue of preexisting input (Rajan et al., 1993; Robert-son and Irvine, 1989; Rajan and Irvine, 1998).

The residue argument predicts that, if all high-CFneurons expressed a new CF as a consequence of residualinputs, then thresholds at the postlesion residual CFwould be higher than CF thresholds in neurons for whichthis frequency was the normal CF, as noted above (seeResults). Furthermore, if all high-CF neurons exhibitedthe same ‘‘new’’ CF, then the residue argument wouldpredict a systematic (although not necessarily monotonic)increase in CF threshold from normal AI across the regioncontaining the expanded representation of a lesion-edgefrequency (Rajan et al., 1993; Rajan and Irvine, 1998). Theobserved progression of CF thresholds in AI differedmarkedly from this prediction from the residue hypothesis(Rajan et al., 1993; Rajan and Irvine, 1998). The same

Fig. 4. Progression of threshold at characteristic frequency (CF)with distance in normal and ipsilesional dorsal cochlear nucleus(DCN). A-D: DCN ipsilateral to the lesioned cochlea in test animals.E-H: Normal (i.e., left) DCN in two control animals (E,F) andcontralesional (i.e., right) DCN in two test animals (G,H). Animalnumber is in the top right corner in A–H. Abscissa layout is that sameas that in Figure 3 and plots distance in DCN relative to the 10-kHzCF point in each penetration in each animal. Positive depths againindicate points farther ventral along the DCN strial axis than the10-kHz CF point, and negative values indicate more dorsally locatedpoints. The ordinate plots threshold at CF, and the shaded regionlabelled NR indicates neuronal clusters that were unresponsive totone-burst stimuli. I–L: Comparison of the threshold-distance progres-sion in normal DCN vs. that in each subcategory of ipsilesional DCN.The full line in each case presents the Beiter-spline of best approxima-tion for the trend of data (symbols) in that plot. I presents thresholddata for the ipsilesional DCN in which the CF-distance plot flattenedout at ,16 kHz (CML 9420, shown in Fig. 3E; CML 95105, shown inFig. 3F). J presents threshold data for the ipsilesional DCN in whichthe CF-distance plot flattened out at ,22 kHz (CML 9418, shown inFig. 3G; CML 9466, shown in Fig. 3H). K presents data for the fournormal DCNs shown in E–H and from the third control animal. Ineach plot, the distance-aligned data from the different animals weresorted; then, the Beiter-spline approximation was fitted by using a lowspline parameter that was kept constant across all three plots (I–K).These fitted lines alone are compared in L, in which the CMLs 16indicates the line fitted in I, CMLs 22 indicates the line fitted in J, andNORMALS indicates the line fitted in K. Note the different abscissascale in L compared with the other plots.

ABSENCE OF PLASTICITY IN ADULT DORSAL COCHLEAR NUCLEUS 43

Figure 5

analysis was applied here. In contrast to AI, in DCN, theCF-threshold progression that was predicted from theresidue argument closely matched the progression of CFthresholds in the lesion animals in which there was aflattening out of the CF-distance plot at about 16 kHz. Italso matched quite well the progression of CF thresholdsin the lesioned animals in which there was a flattening outof the CF-distance plot at about 22 kHz. In the latter case,the mismatch was due to the residue argument underesti-mating the observed increase of CF thresholds in these twoanimals.

These comparisons make it unlikely that any postlesionflattening out of the CF-distance plot in ipsilesional DCNcould account for the effects seen in contralesional AI aftervery similar cochlear lesions. Unlike the AI effects, whichappear to reflect a true plasticity of the tonotopic organiza-tion, the DCN effects appear simply to reflect the residue ofprelesion inputs.

The other major respect in which the postlesion DCNchanges differ from AI changes is that, in the DCN, theflattening out (with or without dips) in the CF-distanceplot was not found uniformly in all penetrations in allanimals. In as many penetrations as those that showedthis effect, the CF-distance plot showed a normal progres-sion until a lesion-edge CF was first encountered and thenno further auditory drive was encountered more dorsally,i.e., deprived DCN regions simply became unresponsive totones. In contrast, in contralesional AI, plasticity effectswere uniform in all penetrations across the tonotopic axisof AI.

The DCN has a highly complex laminar structure bothcytoarchitecturally and in terms of neural connectivity (forreviews, see Cant and Morest, 1984; Moore, 1986; Young etal., 1992), such that it has sometimes been considered tohave a ‘‘cortex-like’’ physiology (cf. Evans, 1968). It was forthis reason that we directed our attention to this subdivi-sion of the cochlear nucleus, because, potentially, it is themost likely to be able to express plasticity of the tonotopicorganization after partial cochlear lesions very similar tothe lesions that cause plasticity in the adult AI (Rajan etal., 1993). The absence of plasticity in our study and thesimilar conclusion drawn by Kaltenbach et al. (1992, 1996)show that having such a complex circuitry does not ensurethat plasticity occurs. Clearly, the DCN must lack neural

mechanisms that are essential for the occurrence of plastic-ity of the type observed in the cortex after cochlear lesions.

The absence of plasticity in the adult DCN is consistentwith the absence of plasticity seen in many (but not all)studies at low subcortical levels in other sensory systems(see above). All of these studies have used partial or totaldeprivation of the outflow from the receptor surface onlyfor relatively brief periods. In contrast, a long period ofsurvival was used in the study by Pons et al. (1991), whofound such extensive cortical plasticity that it was sug-gested that at least some component of this plasticity mustreflect subcortical effects. Although we cannot say whethersuch long survival times would allow plasticity to occur inthe DCN, the survival times after the cochlear lesion inthis study ranged from 2.0 months to 5.5 months, and wehave shown that AI plasticity can be found with a survivaltime of 2.0 months, the shortest used in the AI study(Rajan et al., 1993). Thus, at least, we can say that the AIplasticity expressed with a survival time of 2.0–5.5 monthsafter a partial cochlear lesion in adult animals is not due toplasticity at the DCN. We have also shown (Irvine andRajan, 1994) that similar cochlear lesions and survivaltimes lead to a patchy plasticity in the midbrain inferiorcolliculus (IC), such that some parts of the central nucleusof the IC express plasticity, whereas others show effectsthat are consistent with the residue argument. The pres-ent results also show that, if this patchiness in the IC is areflection that plasticity is expressed in only some of themultiple brainstem pathways to the IC, then it is not dueto inputs from the DCN conveying plastic, tonotopic reorga-nization from this brainstem structure.

LITERATURE CITED

Basbaum, A.I. and P.D. Wall (1976) Chronic changes in the response of cellsin adult cat dorsal horn following partial deafferentation: The appear-ance of responding cells in a previously non-responsive region. BrainRes. 116:181–204.

Cant, N.B. and D.K. Morest (1984) The structural basis for stimulus codingin the cochlear nucleus of the cat. In C.I. Berlin (ed): Hearing Science:Recent Advances. San Diego: College-Hill Press, pp. 373–421.

Chino, Y.M., J.H. Kaas, E.L. Smith, III, A.L. Langton, and H. Cheng (1992)Rapid reorganisation of cortical maps in adult cats following restricteddeafferentation in retina. Vision Res. 32:789–796.

Darian-Smith, C. and C.D. Gilbert (1995) Topographic reorganisation in thestriate cortex of the adult cat and monkey is cortically mediated. J.Neurosci. 15:1631–1647.

Devor, M. and P.D. Wall (1981) Plasticity in the spinal cord sensory mapfollowing peripheral nerve injury in rats. J. Neurosci. 1:679–684.

Evans, E.F. (1968) Upper and lower levels of the auditory system: A contrastof structure and function. In E.R. Caianiello (ed): Neural Networks.Berlin: Springer-Verlag, pp. 24–33.

Eysel, U.T., F. Gonzalez-Aguilar, and U. Mayer (1980) A functional sign ofreorganisation in the visual system of adult cats: Lateral geniculateneurons with displaced receptive fields after lesions of the nasal retina.Brain Res. 181:285–300.

Garraghty, P.E. and J.H. Kaas (1991) Functional reorganisation in adultmonkey thalamus after peripheral nerve injury. Neuroreport 2:747–750.

Gilbert, C.D. and T.N. Weisel (1992) Receptive field dynamics in adultprimary visual cortex. Nature 356:150–152.

Heinen, S.J. and A.A. Skavenski (1991) Recovery of visual responses infoveal V1 neurons following bilateral foveal lesions in adult monkey.Exp. Brain Res. 83:670–674.

Irvine, D.R.F. and R. Rajan (1994) Plasticity of frequency organization ininferior colliculus of adult cats with unilateral restricted cochlearlesions: 17th Midwinter meeting of the Association for Research inOtolaryngology. Assoc. Res. Otolaryngol. 82:S21.

Irvine, D.R.F. and D. Robertson (1990) Reorganization of frequency repre-sentation in auditory cortex of guinea pigs with partial unilateral

Fig. 5. Predictions of the residue hypothesis for the progression ofcharacteristic frequency (CF) thresholds with distance in ipsilesionaldorsal cochlear nucleus (DCN). A–D: Data for the case in which anexpanded CF representation of 16 kHz occurred as a residue ofpreexisting input. E–H: Data for the case in which the expandedresidual CF representation was at 22 kHz. In these plots, data arefrom normal DCN (animal numbers are shown at top right). In A–D,data plotted are the CF thresholds for CFs # 16 kHz; thereafter, thethresholds plotted for clusters with CF . 16 kHz are the thresholds at16 kHz measured in the excitatory response area (ERA). Similarly, inE–H, data plotted are CF thresholds for CFs # 22 kHz; thereafter, thethresholds plotted for clusters with CF . 22 kHz are thresholds at 22kHz measured from the ERA. I,J: Comparison of Beiter-spline lines ofbest approximation with CF-threshold data from normal DCN (NOR-MALS; the same spline fit as in Fig. 4K), from a particular subset ofipsilesional DCN (CML with animal numbers; for I, this is the samespline fit as in Fig. 4I; for J, this is the same spline fit as in Fig. 4J), orfor the CF thresholds predicted from the residual hypothesis for thecase in which an expanded CF representation of 16 kHz (I, PRE-DICTED 16 kHz RESIDUE) or 22 kHz (J, PREDICTED 22 kHzRESIDUE) occurred as a residue of preexisting input. Note thedifferent abscissa scale in I and J compared with the other plots.

ABSENCE OF PLASTICITY IN ADULT DORSAL COCHLEAR NUCLEUS 45

deafness. In M. Rowe and L. Aitkin (eds): Information Processing inMammalian Auditory and Tactile Systems. New York: Wiley-Liss, Inc.,pp. 253–266.

Kaas, J.H., L.A. Krubitzer, Y.M. Chino, A.L. Langton, E.H. Polley, and N.Blair (1990) Reorganization of retinotopic cortical maps in adultmammals after lesions of the retina. Science 248:229–231.

Kaltenbach, J.A., J.M. Czaja, and C.R. Kaplan (1992) Changes in thetonotopic map of the dorsal cochlear nucleus following induction ofcochlear lesions by exposure to intense sound. Hearing Res. 59:213–223.

Kaltenbach, J.A., R.J. Meleca, and P.R. Falzarano (1996) Alterations in thetonotopic map of the cochlear nucleus following cochlear damage. InR.J. Salvi, D. Henderson, F. Fiorino, and V. Colletti (eds): AuditorySystem Plasticity and Regeneration. New York: Thieme Medical Publish-ers, pp. 317–332.

Merzenich, M.M., J.H. Kaas, J.T. Wall, M. Sur, R.J. Nelson, and D.J.Felleman (1983) Topographic reorganization of somatosensory corticalareas 3B and 1 in adult monkeys following restricted deafferentation.Neuroscience 8:33–55.

Merzenich, M.M., R.J. Nelson, M.P. Stryker, M.S. Cynader, A. Schoppman,and J.M. Zook (1984) Somatosensory cortical map changes followingdigit amputation in adult monkeys. J. Comp. Neurol. 224:591–605.

Millar, J., A.L. Basbaum, and P.D. Wall (1976) Restructuring of thesomatotopic map and appearance of abnormal neuronal activity in thegracile nucleus after partial deafferentation. Exp. Neurol. 50:658–672.

Moore, J.K. (1986) Cochlear nuclei: Relationship to the auditory nerve. InR.A. Altschuler, D.W. Hoffman, and R.P. Bobbin (eds): Neurobiology ofHearing: The Cochlea. New York: Raven Press, pp. 283–301.

Pons, T.P., P.E. Garraghty, A.K. Ommaya, J.H. Kaas, E. Taub, and M.Mishkin (1991) Massive cortical reorganisation after sensory deafferen-tation in adult macaques. Science 252:1857–1860.

Pubols, L.M. and G.L. Brenowitz (1982) Maintenance of dorsal hornsomatotopic organization and increased high-threshold response aftersingle-root or spared-root deafferentation in cats. J. Neurophysiol.47:103–112.

Pubols, L.M. and M.E. Goldberger (1980) Recovery of function in dorsalhorn following partial deafferentation. J. Neurophysiol. 43:102–117.

Rajan, R. (1997) Auditory response areas altered by intermodulationdistortion products from background tones. Neuroreport 8:3177–3182.

Rajan, R. and D.R.F. Irvine (1996) Features of, and boundary conditions for,lesion-induced reorganisation of adult auditory cortical maps. In R.J.Salvi, D. Henderson, F. Fiorino, and V. Colletti (eds): Auditory SystemPlasticity and Regeneration. New York: Thieme Medical Publishers, pp.224–237.

Rajan, R. and D.R.F. Irvine (1998) Neuronal responses across cortical fieldA1 in plasticity induced by peripheral auditory damage. Audiol. Neuro-Otol. 3:123–144.

Rajan, R., D.R.F. Irvine, and J.F. Cassell (1991) Normative N1 audiogramdata for the barbiturate-anaesthetised domestic cat. Hearing Res.53:153–158.

Rajan, R., D.R.F. Irvine, L.Z. Wise, and P. Heil (1993) Effect of unilateralpartial cochlear lesions in adult cats on the representation of lesionedand unlesioned cochleas in primary auditory cortex. J. Comp. Neurol.338:17–49.

Rasmusson, D. (1982) Reorganization of raccoon somatosensory cortexfollowing removal of the fifth digit. J. Comp. Neurol. 205:313–326.

Rasmusson, D. (1988) Projections of digit afferents to the cuneate nucleusin the raccoon before and after partial deafferentation. J. Comp. Neurol.277:549–556.

Rasmusson, D. (1996) Changes in the organization of the ventroposteriorlateral thalamic nucleus after digit removal in adult raccoon. J. Comp.Neurol. 364:92–103.

Robertson, D. and D.R.F. Irvine (1989) Plasticity of frequency organizationin auditory cortex of guinea pigs with partial unilateral deafness. J.Comp. Neurol. 282:456–471.

Schmid, L.M., M.G. Rosa, M.B. Calford, and J.S. Ambler (1996) Visuotopicreorganisation in the primary visual cortex of adult cats followingmonocular and binocular retinal lesions. Cerebral Cortex 6:388–405.

Schwaber, M.K., P.E. Garraghty, and J.H. Kaas (1993) Neuroplasticity ofthe adult primate auditory cortex following cochlear hearing loss. Am.J. Otol. 14:252–258.

Spirou, G.A., B.J. May, D.D. Wright, and D.K. Ryugo (1993) Frequencyorganization of the dorsal cochlear nucleus in cats. J. Comp. Neurol.329:36–52.

Young, E.D., G.A. Spirou, J.J. Rice, and H.F. Voigt (1992) Neural organiza-tion and responses to complex stimuli in the dorsal cochlear nucleus.Phil. Trans. R. Soc. London B 336:407–413.

46 R. RAJAN AND D.R.F. IRVINE