Journal of Neurophysiology

Adaptive Adjustment of Connectivity in the Inferior Colliculus Revealed by Focal Pharmacological Inactivation

Joshua I. Gold, Eric I. Knudsen

Abstract

In the midbrain sound localization pathway of the barn owl, a map of auditory space is synthesized in the external nucleus of the inferior colliculus (ICX) and transmitted to the optic tectum. Early auditory experience shapes these maps of auditory space in part by modifying the tuning of the constituent neurons for interaural time difference (ITD), a primary cue for sound-source azimuth. Here we show that these adaptive modifications in ITD tuning correspond to changes in the pattern of connectivity within the inferior colliculus. We raised owls with an acoustic filtering device in one ear that caused frequency-dependent changes in sound timing and level. As reported previously, device rearing shifted the representation of ITD in the ICX and tectum but not in the primary source of input to the ICX, the central nucleus of the inferior colliculus (ICC). We applied the local anesthetic lidocaine (QX-314) iontophoretically in the ICC to inactivate small populations of neurons that represented particular values of frequency and ITD. We measured the effect of this inactivation in the optic tecta of a normal owl and owls raised with the device. In the normal owl, inactivation at a critical site in the ICC eliminated responses in the tectum to the frequency-specific ITD value represented at the site of inactivation in the ICC. The location of this site was consistent with the known pattern of ICC-ICX-tectum connectivity. In the device-reared owls, adaptive changes in the representation of ITD in the tectum corresponded to dramatic and predictable changes in the locations of the critical sites of inactivation in the ICC. Given that the abnormal representation of ITD in the tectum depended on frequency and was likely conveyed directly from the ICX, these results suggest that experience causes large-scale, frequency-specific adjustments in the pattern of connectivity between the ICC and the ICX.

INTRODUCTION

Early experience can have a profound influence on the representation of sensory information in the brain. This influence is thought to involve plasticity in the connectivity between neurons, which can shape their stimulus selectivity and other complex response properties (Buonomano and Merzenich 1998). A prerequisite for understanding the mechanisms responsible for this kind of plasticity is the identification of sites in the nervous system where experience-induced changes in connectivity take place. Here we use the technique of reversible, focal inactivation of neurons with the anesthetic lidocaine (QX-314) to identify auditory experience-induced changes in functional connectivity in the midbrain auditory localization pathway of the barn owl.

The midbrain of birds and mammals contains a map of auditory space that is based on the tuning of neurons for sound localization cues, such as interaural time differences (ITDs) and interaural level differences (ILDs) (Carlile and King 1994; Olsen et al. 1989; Wise and Irvine 1985). In the barn owl, this map is created by merging information about cues across frequency channels to eliminate ambiguities that are inherent to individual cues (Brainard et al. 1992). This merging of information occurs in the projection from the tonotopically organized central nucleus of the inferior colliculus (ICC) to its space-mapped external nucleus (ICX).

The map of auditory space in the ICX is shaped by experience. The shaping influence of experience on the space map is typically assessed in the optic tectum (also called the superior colliculus), which receives the map by a topographic projection from the ICX. For example, raising owls with abnormal auditory experience has been shown to cause adaptive adjustment of the tectal space map (Gold and Knudsen 1999; Knudsen 1983, 1985). This adjustment corresponds to changes in unit tuning for ITD and ILD in both the ICX and tectum (Gold and Knudsen 2000a,b; Mogdans and Knudsen 1992, 1993).

The same auditory manipulations that alter auditory tuning in the ICX and tectum have little or no effect on the representations of binaural cues in the ICC (Gold and Knudsen 2000b; Mogdans and Knudsen 1994). Because the ICC projects directly to the ICX, these data suggest that plasticity occurs in the pattern of ICC-ICX connectivity (Fig. 1). Indeed, visual experience-induced adjustment of the space map in the ICX (and tectum) is associated with systematic changes in the anatomical projection from the ICC to the ICX (Feldman and Knudsen 1997).

Fig. 1.

The ascending interaural time difference (ITD) pathway to the optic tectum (adapted from Brainard and Knudsen 1993). All nuclei are bilaterally symmetrical. Vertical tick marks indicate tonotopic organization, found in the nucleus magnocellularis (NM), nucleus laminaris (NL, the 1st site of interaural phase comparison) (Carr and Konishi 1988, 1990; Sullivan and Konishi 1986), and both the core and lateral shell subdivisions of the central nucleus of the inferior colliculus (ICC). The representation of ITD is normal in the ICC of owls raised with abnormal vision (Brainard and Knudsen 1993) or hearing (Gold and Knudsen 2000b). In contrast, both auditory and visual experience shape the representations of ITD in the external nucleus of the inferior colliclus (ICX) and the optic tectum, where neurons are organized with respect to their auditory spatial tuning to form physiological maps of auditory space (Knudsen 1982;Knudsen and Konishi 1977, 1978). We tested whether the influence of abnormal auditory experience on these maps resulted from plasticity in the projection from the lateral shell of the ICC to the ICX.

In this study, we used focal pharmacological inactivation in the ICC to test whether changes in the midbrain space map that are induced by abnormal auditory experience are accompanied by changes in ICC-ICX connectivity. We found that auditory experience-induced changes in unit tuning for ITD in the tectum (and equivalent changes in the ICX) (Gold and Knudsen 2000b) corresponded to changes in the locations of the ICC units that drove the tectal responses. The results indicate that changes in functional connectivity within the inferior colliculus can account for the adaptive adjustment of the midbrain auditory space maps.

METHODS

Auditory experience

We raised four barn owls (Tyto alba), three with an acoustic filtering device implanted chronically in the right ear to alter auditory experience and one without a device. The device was a custom-designed plastic chamber that was sutured in the right ear canal and rested just behind the preaural flap and in front of the facial ruff feathers. A more detailed description of the device, including cochlear microphonic measurements of its frequency-specific effects on sound timing and level, can be found in Gold and Knudsen (1999). All four owls were used in previous studies (Gold and Knudsen 1999, 2000a,b).

The owls were initially raised in brooding boxes with their siblings. At ∼25 days of age, the owls were anesthetized with halothane (1%) in a mixture of oxygen and nitrous oxide (5:4) and dense-foam-rubber earplugs (E.A.R. Cabot) were sutured into both ear canals. The earplugs, which attenuate frequencies between 2 and 8 kHz by ∼20–40 dB (Knudsen et al. 1984), were used to limit auditory experience while the owls' ear canals were open but not yet large enough to accommodate the acoustic device. At ∼35 days of age, the owls were re-anesthetized, the binaural foam plugs were removed, and the acoustic device was sutured into the right ear canal. Each owl was then placed in an individual cage located next to a large flight cage that housed many adult owls, providing a rich visual and auditory environment. When the owls could fly, at ∼60 days of age, they were placed in the large flight cage.

The owls were provided for in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the guidelines of the Stanford University Institutional Animal Care and Use Committee.

Electrophysiology

Owls were prepared for electrophysiological measurements as described in detail elsewhere (Gold and Knudsen 2000b). Briefly, each owl was anesthetized with halothane (1%) in a mixture of oxygen and nitrous oxide (5:4), a stainless steel plate was cemented to the base of the skull, and small craniotomies were made. All incised tissues were infused with lidocaine, treated with betadine, and sutured back together. To begin each recording session, the owl was anesthetized with halothane and nitrous oxide, wrapped in a soft leather jacket, and given an intramuscular injection of 2.5% dextrose in sterile saline. The acoustic device was removed, and the eardrum and ear canal were inspected for damage and cleaned of earwax. The owl was suspended in a prone position in a stereotaxic apparatus located in a sound-attenuated chamber. The owl's head was held in place with the surgically implanted steel plate and positioned using retinal landmarks.

Unit activity was recorded extracellularly with tungsten microelectrodes (1–2 MΩ at 1 kHz). A level discriminator was used to isolate action potentials (“spikes”) generated by a small number of nearby neurons. For each sound presentation, the response was calculated as the number of spikes in the 100 ms immediately following stimulus onset minus the number of spikes in the 100 ms immediately preceding stimulus onset.

Nitrous oxide was normally administered continuously through the course of the experiment. At the end of a session, the craniotomy was bathed in chloramphenicol (0.5%) and re-sealed with dental acrylic, and the device was sutured back into place. The owl was placed under a heat lamp and monitored until fully recovered (∼1–2 h).

Dichotic stimulation

Dichotic stimuli consisted of digitally filtered broadband (3–12 kHz passband) and narrowband (1-kHz bandwidth centered on the given frequency) noise bursts, 50 ms in duration, with 0- and 5-ms rise/fall times, respectively. Sounds were presented via matched Knowles earphones (ED-1914) coupled to damping assemblies (BF-1743), placed ∼5 mm from each tympanic membrane. The frequency response of each earphone was flat to within ±2 dB from 3 to 10 kHz. For a given ITD, ILD, or frequency tuning curve, the peak width was defined as the range of stimulus values that elicited a response >50% of the maximum response. The best value was defined as the center of this range. Positive and negative ITD values refer to right- and left-ear leading, respectively.

Iontophoresis protocol

Iontophoresis electrodes were made by pulling a five-barrel glass capillary (WPI) that had a 7-μm carbon fiber (Amoco Performance Products, Alpharetta, GA) inserted into the central barrel (Armstrong-James and Millar 1979). The tips of the electrode were cut so that the inner diameter of each outer barrel was 1–3 μm and the tips were flush with the carbon fiber. The central barrel was filled with 1 M NaCl to make contact with the carbon fiber for recording. A chloride-coated silver wire made contact with the solution in each barrel. QX-314 (Research Biochemicals), a quaternary lidocaine derivative, was used to block fast, Na+-dependent action potentials (Connors and Prince 1982) in the ICC. Two of the four outer barrels of the iontophoresis electrode were filled with 2% QX-314 solution in dH2O (58 mM), with HCl, NaOH, or both added to achieve a pH of 6.5. The lidocaine solution was applied by iontophoresis (NeuroData IPX-5), typically using 90-nA injection currents through each of two barrels. Before and after injections, a 10-nA retaining current was applied to the lidocaine-filled barrels.

Experimental protocol

ITD is represented systematically along the rostrocaudal axes of the ICC, ICX, and tectum as described in Fig.2 A. We recently demonstrated that device rearing causes adaptive shifts in the representation of ITD in both the optic tectum and the ICX (Gold and Knudsen 2000a,b). In contrast, device rearing does not affect the representation of ITD in the ICC. As shown in Fig. 2 B, changes in the pattern of ICC-ICX connectivity could, in principle, account for these device-induced shifts. According to this model, the ICX-tectum connections remain unchanged, and the adaptive response properties created in the ICX are conveyed along normal projections to the optic tectum.

Fig. 2.

Model of plasticity in ICC-ICX connectivity to account for device-induced changes in ITD tuning. The organization of the lateral shell of the ICC is identical in normal (Wagner et al. 1987) and device-reared (Gold and Knudsen 2000b) owls: neurons in an individual lamina perpendicular to the dorsoventral axis of the nucleus are narrowly tuned to a single frequency, the value of which increases systematically with depth. In addition, a dorsoventral array of ICC neurons that spans these frequency laminae represents a single ITD, the value of which varies systematically from 0 μs rostrally to more contralateral-ear leading values more caudally. A: in normal owls, an array of ICC neurons representing a single ITD and a broad range of frequencies projects to an ICX neuron that, accordingly, is tuned to the same ITD (0 μs in the figure) and frequency range. That neuron, in turn, projects to a neuron in the tectum, conferring tuning to the same ITD and broad frequency range (Knudsen and Knudsen 1983). As in the ICC, neurons in both the ICX and the tectum are organized with respect to the value of ITD they represent, with 0 μs represented rostrally and more contralateral-ear leading values represented more caudally (Moiseff and Konishi 1981; Olsen et al. 1989; Wagner et al. 1987). B: in device-reared owls, ITD tuning in the ICX and tectum is shifted from normal by ∼50–80 μs toward open-ear leading for 4-kHz stimuli and ∼15 μs toward affected-ear leading for 8-kHz stimuli with a loss of responsiveness to 6-kHz stimuli (Gold and Knudsen 2000a,b). For a neuron normally tuned to 0-μs ITD in the right ICX (or tectum), these device-induced changes could result from the loss of the normal ICC-ICX connections (dashed arrows), and the acquisition of input to the ICX from a more caudally located site in the 4-kHz lamina and a more rostrally located site in the 8-kHz lamina of the ICC (solid arrows).

We tested this model by recording from neurons in the tectum while simultaneously inactivating local regions of the ICC in both normal and device-reared owls. The idea was to test whether device rearing changed the locations of the ICC units that drove the tectal responses. We measured the effects of ipsilateral ICC inactivation on responses in the tectum and not the ICX because it was easier to place electrodes simultaneously in the ICC and tectum than in the ICC and ICX.

The optic tectum was targeted stereotaxically and was recognized by characteristic bursting activity and spatially restricted auditory and visual receptive fields (RFs) (Knudsen 1982). Electrode position within the tectum was determined on the basis of visual RF location, which was used to determine predicted normal ITD values (Olsen et al. 1989). The inferior colliculus was targeted stereotaxically by positioning the electrode relative to the tectal representation of frontal space. Within the inferior colliculus, the lateral shell of the ICC was recognized using a number of criteria, including stereotaxic position, response latencies, sensitivity for interaural level difference (ILD), width of frequency tuning, and progression of best frequency with dorsoventral depth (Gold and Knudsen 2000b).

With the injection electrode positioned in the lateral shell, a tungsten electrode was positioned in the ipsilateral tectum. In most cases, we targeted tectal sites with a best ITD and best ILD that matched the best values at the ICC site, using a narrowband stimulus with a center frequency that matched the best frequency at the ICC site (typically ∼4 kHz). We also measured ITD tuning at the tectal recording site using a narrowband stimulus with a center frequency that did not match the best frequency at the ICC site (typically ∼8 kHz). During drug injection, responses at the ICC injection site to a broadband stimulus, using the site's best ITD and best ILD, were measured every 3–6 s until they were <35% of preinjection levels (∼5–15 min). A series of ITD tuning curves of 10–20 repetitions each and alternately using the 4- or 8-kHz stimulus was then measured at the tectal recording site. Finally, a retaining current was turned on or the injection electrode was removed, and more ITD tuning curves were measured at the tectal recording site as the ICC recovered from the lidocaine injection.

RESULTS

We used focal pharmacological inactivation to test whether changes in the pattern of connectivity in the inferior colliculus were responsible for auditory experience-induced changes in the midbrain ITD map. Below, we first show that injection of lidocaine in the ICC reversibly blocked auditory-evoked activity at the injection site. Next, we demonstrate that, consistent with the pattern of connectivity in normal owls, inactivation at a critical site in the ICC blocked ITD tuning at a site in the tectum in a frequency-dependent manner. Finally, we show that the locations of these critical, frequency-specific ICC sites were altered by auditory experience.

Local effect of lidocaine in the ICC

Lidocaine reversibly blocked auditory-evoked responses at the ICC injection site. Figure 3 illustrates this effect at a site in the lateral shell of the ICC of a device-reared owl. Before lidocaine application, units at this site were tuned to 3.9 kHz (data not shown) and −82 μs ITD (Fig. 3, B andC). Ten minutes of lidocaine injection eliminated auditory responses at this site (Fig. 3, B and D). One hundred thirty minutes after drug injection ceased, unit responses returned to preinjection levels (Fig. 3, B andE). These results were typical of the effect of lidocaine on responses at the ICC injection site in both normal and device-reared owls: 5–15 min of lidocaine injection caused auditory responses to a broadband stimulus at the site's best ITD to be reduced by ≥70% relative to the baseline level (Fig. 4). For all 12 sites tested, responses returned to ≥90% of baseline levels within 150 min of termination of drug injection.

Fig. 3.

Local effect of lidocaine injection on auditory responses at a site in the lateral shell of the ICC. A: schematic of the placement of the electrode in the ICC for this experiment (see Fig. 2).B: ITD tuning curves (data are means ± SE,n = 20–50 repetitions) measured in the ICC using a broadband stimulus, before (baseline), during (+ lidocaine), and 130 min after (recovery) a 10-min injection of anesthetic at the recording site. *, ITDs for which responses during the injection of anesthetic were significantly different from baseline (unpairedt-test, P < 0.01).C–E: peristimulus time histograms of responses to 30 repetitions of a broadband stimulus near the site's best ITD (−80 μs) and best ILD (−5 dB), measured before (C), during (D), and after recovery from (E) injection of anesthetic into the ICC. The broadband stimulus was presented for 50 ms, starting at time 0 (black bars).

Fig. 4.

Summary of the local effect of lidocaine injection on auditory responses in the lateral shell of the ICC. Mean responses to a broadband stimulus presented at the site's best ITD and best ILD after 5–15 min of lidocaine injection are plotted as a function of baseline responses. Circles and error bars represent means ± SE for individual sites (n = 20–50 repetitions); error bars in most cases are obscured by the symbols. Symbols to the right of the dashed line indicate a drug-induced reduction of responses. In all cases, this reduction was significant.

Effect on tectal responses in normal owls

Units in the optic tectum are tuned for ITD but, unlike units in the ICC, are broadly tuned for frequency. Thus we used narrowband stimuli that matched the frequency tuning at the ICC injection sites to test the effects of ICC inactivation on frequency-specific ITD representations in the tectum.

Figure 5 illustrates responses to binaural stimuli at a tectal site before, during, and after recovery from lidocaine injection in the ICC. The injection site was located at the rostral end of the lateral shell in the ICC (Fig. 5 A). The units at this site were tuned to 5.1 kHz and 5 μs ITD. The recording site was located at the rostral end of the optic tectum. Prior to the injection of lidocaine in the ICC, the units at this site responded robustly to a narrowband stimulus centered on 5.1 kHz, with a best ITD of 9 μs (Fig. 5, B and C). This stimulus and ITD tuning matched the frequency and ITD tuning, respectively, at the ICC injection site. After 9 min of lidocaine injection, responses at the ICC injection site were blocked (data not shown). Responses at the tectal recording site to the 5.1-kHz stimulus were blocked, as well (Fig. 5, B and D). The tectal responses returned to baseline levels 45 min after termination of the drug injection (Fig. 5, B and E). Similar results were obtained in two other experiments in which the best ITD at the ICC injection site was within 11 μs of the frequency-matched best ITD at the tectal recording site.

Fig. 5.

Effect of lidocaine injection in the ICC on ITD- and frequency-matched responses in the optic tectum of a normal owl. A: schematic of the placement of the drug-injection electrode in the ICC and the recording electrode in the tectum for this experiment. Arrows indicate the connectivity that can account for the frequency-specific ITD tuning at the tectal site (see Fig. 2). B: ITD tuning curves (data are means ± SE, n = 30–60 repetitions) measured in the optic tectum before (baseline), during (+ lidocaine), and 45 min after (recovery) a 9-min injection of anesthetic into the lateral shell of the ICC. At the site of injection in the ICC, the best frequency was 5.1 kHz (width = 1.3 kHz) and the best ITD was 5 μs (▴). The ITD tuning curves were measured using a narrowband stimulus that matched the frequency tuning at the injection site (passband = 4.7–5.5 kHz). *, ITDs for which responses during the injection of anesthetic were significantly different from baseline (unpaired t-test,P < 0.01). C–E: peristimulus time histograms of responses to 60 repetitions of a broadband stimulus near the injection site's best ITD (10 μs), measured before (C), during (D), and after recovery from (E) injection of anesthetic into the ICC. The broadband stimulus was presented for 50 ms, starting at time 0(black bars).

In contrast, responses at the same tectal site to a narrowband stimulus centered on 7.5 kHz, instead of 5.1 kHz, were not affected by injection of lidocaine at the same ICC site (Fig.6). This stimulus did not match the frequency tuning at the injection site. Prior to the injection of lidocaine in the ICC, the tectal units responded to the 7.5-kHz stimulus with a best ITD of 11 μs. After 23 min of lidocaine injection in the ICC, responses at the tectal recording site were unchanged relative to baseline levels.

Fig. 6.

Effect of lidocaine injection in the ICC on frequency-unmatched responses in the optic tectum of a normal owl. A: schematic of the placement of the drug-injection electrode in the ICC and the recording electrode in the tectum for this experiment, which are the same sites as in Fig. 5. Arrows indicate the connectivity that can account for the frequency-specific ITD tuning at the tectal site (see Fig. 2). B: ITD tuning curves (data are means ± SE, n = 30–60 repetitions) measured in the optic tectum before (baseline), during (+ lidocaine), and 45 min after (recovery) a 13-min injection of anesthetic into the lateral shell of the ICC. At the site of injection in the ICC, the best frequency was 5.1 kHz (width = 1.3 kHz) and the best ITD was 5 μs. The ITD tuning curves were measured using a narrowband stimulus that did not match the frequency tuning at the injection site (passband = 7.1–7.9 kHz). Responses during and after lidocaine injection were not significantly different from baseline at all ITDs.

Unit responses in the tectum were also unaffected by injection of lidocaine into the ICC when the ITD tuning of units at the ICC injection site did not match that of the units at the tectal recording site. For the experiment illustrated in Fig.7, the injection site was located at the rostral end of the lateral shell in the ICC. The units at this site were tuned to 4.8 kHz and −15 μs ITD. The recording site was located at the caudal end of the optic tectum. Prior to the injection of lidocaine in the ICC, the units at this site responded to a narrowband stimulus centered on 4.8 kHz with a best ITD of −66 μs. After 25 min of lidocaine injection, responses at the ICC injection site were blocked (data not shown), but there was no effect on the responses at the tectal recording site to the 4.8-kHz stimulus at any ITD. Similar results were obtained in one other experiment in which the frequency-matched best ITDs at the ICC injection site and the tectal recording site differed by 36 μs.

Fig. 7.

Effect of lidocaine injection in the ICC on ITD-unmatched responses in the optic tectum of a normal owl. A: schematic of the placement of the drug-injection electrode in the ICC and the recording electrode in the tectum for this experiment. Arrows indicate the connectivity that can account for the frequency-specific ITD tuning at the tectal site (see Fig. 2). B: ITD tuning curves (data are means ± SE, n = 15–30 repetitions) measured in the optic tectum before (baseline), during (+ lidocaine), and 35 min after (recovery) a 25-min injection of anesthetic into the lateral shell of the ICC. At the site of injection in the ICC, the best frequency was 4.8 kHz and the best ITD was −15 μs (▴). The ITD tuning curves were measured using a narrowband stimulus that matched the frequency tuning at the injection site (passband = 4.7–4.9 kHz). Responses during and after lidocaine injection were not significantly different from baseline at all ITDs.

Thus in normal owls, auditory responses in the optic tectum were reversibly eliminated by injection of lidocaine into the ICC only for those responses that corresponded to the frequency-specific ITD tuning at the site of drug injection in the ICC. Tectal responses to other frequencies and ITDs were not affected.

Effect on tectal responses in device-reared owls

Device rearing causes frequency-dependent changes in the representation of ITD in the optic tectum (Gold and Knudsen 2000a). These changes include shifts in ITD tuning of ∼50–80 μs toward open-ear leading for narrowband stimuli near 4 kHz. In contrast, device rearing does not affect the representation of ITD in the ICC, with contralateral-ear leading ITD values identically represented in the lateral shell of the ICC in normal and device-reared owls (Gold and Knudsen 2000b). For this study, we focused on device-induced changes in the 4-kHz ITD tuning of neurons in the rostral portion of the right tectum. Neurons at this location typically have a 4-kHz best ITD of ∼0 μs in normal owls but may be tuned to −80 μs in owls raised with the device in the right ear. The advantage of focusing on this particular aspect of the adaptive change is that both the normal and learned ITD values are represented in well-characterized regions of the ipsilateral ICC (see Fig. 2) (Brainard and Knudsen 1993; Gold and Knudsen 2000b; Wagner et al. 1987).

Figure 8 illustrates responses to binaural stimuli at a tectal site before, during, and after lidocaine injection in the ICC. The injection site was located at the caudal end of the lateral shell in the ICC (Fig. 8 A). The units at this site were tuned to 3.9 kHz and −98 μs ITD. The recording site was located at the rostral end of the optic tectum. Prior to the injection of lidocaine in the ICC, the units at this site responded robustly to a narrowband stimulus centered on 4.0 kHz with a best ITD of −84 μs, which is ∼80 μs more open-ear leading than normal (Fig. 8,B and C). This 4.0-kHz stimulus and shifted ITD tuning matched the frequency and ITD tuning, respectively, at the ICC injection site. After 25 min of lidocaine injection, responses at the ICC injection site were blocked (data not shown). After 50 min of injection, responses at the tectal recording site to the 4.0-kHz stimulus were significantly reduced as well (Fig. 8, B andD). The tectal responses returned to baseline levels 30 min after termination of drug injection (Fig. 8, B andE).

Fig. 8.

Effect of lidocaine injection in the ICC on ITD- and frequency-matched responses in the optic tectum of a device-reared owl. A: schematic of the placement of the drug-injection electrode in the ICC and the recording electrode in the tectum for this experiment. Dashed and solid arrows indicate the connectivity that can account for the normal and shifted frequency-specific ITD tuning, respectively, at the tectal site (see Fig. 2). B: ITD tuning curves (data are means ± SE, n = 35–40 repetitions) measured in the optic tectum before (baseline), during (+ lidocaine), and 30 min after (recovery) a 50-min injection of anesthetic into the lateral shell of the ICC. At the site of injection in the ICC, the best frequency was 3.9 kHz (width = 1.1 kHz) and the best ITD was −98 μs (▴). The ITD tuning curves were measured using a narrowband stimulus that matched the frequency tuning at the injection site (passband = 3.5–4.5 kHz). *, ITDs for which responses during the injection of anesthetic were significantly different from baseline (unpaired t-test,P < 0.01). C–E: peristimulus time histograms of responses to 20 repetitions of a broadband stimulus near the injection site's best ITD (−95 μs), measured before (C), during (D), and after recovery from (E) injection of anesthetic into the ICC. The broadband stimulus was presented for 50 ms, starting at time 0(black bars).

In contrast, responses at the same tectal site to a narrowband stimulus centered on 8.0 kHz, instead of 4.0 kHz, were not blocked by injection of lidocaine at the same ICC site (Fig.9). This stimulus did not match the frequency tuning at the injection site. Before injection of lidocaine, the units at the tectal site responded to the 8.0-kHz stimulus with a best ITD of −96 μs. After 28 min of lidocaine injection in the ICC, responses at the tectal recording site were greater than baseline levels. These responses returned to baseline levels 50 min after termination of drug injection.

Fig. 9.

Effect of lidocaine injection in the ICC on frequency-unmatched responses in the optic tectum of a device-reared owl. A: schematic of the placement of the drug-injection electrode in the ICC and the recording electrode in the tectum for this experiment, which are the same sites as in Fig. 8. Arrows indicate the connectivity that can account for the normal frequency-specific ITD tuning at the tectal site (see Fig. 2). B: ITD tuning curves (data are means ± SE, n = 10–30 repetitions) measured in the optic tectum before (baseline), during (+ lidocaine), and 50 min after (recovery) a 50-min injection of anesthetic into the lateral shell of the ICC. At the site of injection in the ICC, the best frequency was 3.9 kHz (width = 1.1 kHz) and the best ITD was −98 μs. The ITD tuning curves were measured using a narrowband stimulus that did not match the frequency tuning at the injection site (passband = 7.5–8.5 kHz). Asterisks indicate ITDs for which responses during the injection of anesthetic were significantly different from baseline (unpaired t-test,P < 0.01).

Unit responses in the tectum were also unaffected by injection of lidocaine into the ICC when the ITD tuning of units at the ICC injection site matched the normal, but not the shifted, ITD tuning of the units at the tectal recording site. For the experiment illustrated in Fig. 10, the injection site was located at the rostral end of the lateral shell in the ICC. The units at this site were tuned to 4.5 kHz and 0 μs ITD. The recording site was located at the rostral end of the optic tectum. Prior to the injection of lidocaine in the ICC, the units at this site, which normally would have responded to a narrowband stimulus centered on 4.5 kHz with a best ITD of ∼0 μs, were instead tuned to −68 μs ITD. After 25 min of lidocaine injection, responses at the ICC injection site were blocked (data not shown), but there was no effect on the responses at the tectal recording site to the 4.5-kHz stimulus at any ITD.

Fig. 10.

Effect of lidocaine injection in the ICC on ITD-unmatched responses in the optic tectum of a device-reared owl. A: schematic of the placement of the drug-injection electrode in the ICC and the recording electrode in the tectum for this experiment. Dashed and solid arrows indicate the connectivity that can account for the normal and shifted frequency-specific ITD tuning, respectively, at the tectal site (see Fig. 2). B: ITD tuning curves (data are means ± SE, n = 50–100 repetitions) measured in the optic tectum before (baseline), during (+ lidocaine), and 35 min after (recovery) a 25-min injection of anesthetic into the lateral shell of the ICC. At the site of injection in the ICC, the best frequency was 4.5 kHz and the best ITD was 0 μs (▴). The ITD tuning curves were measured using a narrowband stimulus that matched the frequency tuning at the injection site (passband = 4.3–4.7 kHz). Responses during and after lidocaine injection were not significantly different from baseline at all ITDs.

Figures 11 and12 summarize data from seven experiments in device-reared owls. In each experiment, the injection site was located in the ∼4 kHz frequency lamina at the caudal end of the ICC. The recording site was located at the rostral end of the optic tectum and was tuned to a shifted, ∼4-kHz ITD that matched the ITD tuning at the ICC injection site. As shown in Fig. 11, 15–90 min of lidocaine injection in the ICC in all cases caused a significant reduction of responses in the optic tectum for a stimulus that matched the frequency and ITD tuning at the ICC injection site (unpairedt-test, P < 0.05). In all but one case (Fig. 11 E), responses during lidocaine injection were also significantly less than responses after termination of the injection. As shown in Fig. 12 A, these effects were confined to stimulus ITDs that matched the best ITD at the ICC injection site to within 10 μs. In contrast, responses to an ∼8-kHz stimulus that did not match the frequency tuning at the ICC injection site were not substantially affected by lidocaine injection at any ITD (Figs. 11 and12 B).

Fig. 11.

Summary of the effect of lidocaine injection in the ICC on responses in the optic tectum of device-reared owls. For each experiment (individual panels), the ITD tuning at the ICC injection site matched the frequency-specific ITD tuning at the tectal recording site. ● and ○, responses measured using a narrowband stimulus that matched the frequency and ITD tuning at the ICC injection site. ▴ and ▵, responses measured using a narrowband stimulus that did not match the frequency tuning at the ICC injection site. Symbols and error bars represent means ± SE from blocks of measurements using a particular stimulus (n = 20–100 repetitions). ● and ▴, responses were significantly different from baseline (unpaired t-test, P < 0.05), where baseline was averaged over all preinjection responses using the same stimulus. Black bars indicate the time of lidocaine injection in the ICC. A: ICC injection site: best ITD = −76 μs, best frequency = 3.9 kHz. Tectal recording site: matched stimulus (3.6–4.2 kHz, ● and ○), best ITD = −89 μs, stimulus ITD = −80 μs; unmatched stimulus (7.6–8.2 kHz, ▵), best ITD = −116 μs, stimulus ITD = −120 μs. B: ICC injection site: best ITD = −98 μs, best frequency = 3.9 kHz. Tectal recording site: matched stimulus (3.5–4.5 kHz), best ITD = −84 μs, stimulus ITD = −95 μs; unmatched stimulus (7.5–8.5 kHz), best ITD = −96 μs, stimulus ITD = −95 μs. C: ICC injection site: best ITD = −93 μs, best frequency = 4.3 kHz. Tectal recording site: matched stimulus (4.2–4.4 kHz), best ITD = −101 μs, stimulus ITD = −95 μs; unmatched stimulus (7.9–8.1 kHz), best ITD = −125 μs, stimulus ITD = −110 μs.D: ICC injection site: best ITD = −105 μs, best frequency = 4.1 kHz. Tectal recording site: matched stimulus (3.5–4.5 kHz), best ITD = −94 μs, stimulus ITD = −100 μs. E: ICC injection site: best ITD = −113 μs, best frequency = 3.8 kHz. Tectal recording site: matched stimulus (3.5–4.5 kHz), best ITD = −115 μs, stimulus ITD = −115 μs; unmatched stimulus (7.5–8.5 kHz), best ITD = −124 μs, stimulus ITD = −130 μs. F: ICC injection site: best ITD = −100 μs, best frequency = 3.9 kHz. Tectal recording site: matched stimulus (3.7–4.7 kHz), best ITD = −99 μs, stimulus ITD = −110 μs; unmatched stimulus (6.5–7.5 kHz), best ITD = −124 μs, stimulus ITD = −125 μs.G: ICC injection site: best ITD = −122 μs, best frequency = 3.9 kHz. Tectal recording site: matched stimulus (3.5–4.5 kHz), best ITD = −126 μs, stimulus ITD = −130 μs; unmatched stimulus (7.0–8.0 kHz), best ITD = −138 μs, stimulus ITD = −140 μs.

Fig. 12.

Effect of lidocaine injection in the ICC on responses to different ITDs at a matched recording site in the optic tectum. The abscissa represents the stimulus ITD relative to the best ITD at the ICC injection site. Only those ITDs that elicited ≥25% of the maximum response using that stimulus during baseline collection are included. The ordinate represents an index of the effect of lidocaine injection, computed as (response with lidocaine in the ICC − response without lidocaine in the ICC)/(response with lidocaine in the ICC + response without lidocaine in the ICC). “Response with lidocaine in the ICC” refers to data from the last block of measurements taken during drug application (see Fig. 11). “Response without lidocaine in the ICC” refers to combined baseline (all preinjection measurements) and recovery (the last block of measurements taken after drug offset). ● and ▴, responses with lidocaine in the ICC were significantly different from both baseline and recovery responses (unpaired t-test, P < 0.05) but baseline and recovery responses were not different from each other.A: data collected using a ∼4-kHz stimulus that matched the frequency tuning at the ICC injection site. Note that significant effects were measured only for stimulus ITDs that matched the ICC injection site best ITD to within 10 μs (shaded region).B: data collected using a ∼8-kHz stimulus that did not match the frequency tuning at the ICC injection site.

Figure 13 summarizes the effect of lidocaine injection in the ICC on auditory responses in the optic tectum of normal and device-reared owls. In general, responses in the tectum were substantially reduced following injection of the anesthetic in the ICC only when the frequency and ITD tuning at the ICC injection site matched the frequency-specific ITD tuning at the tectal recording site (Fig. 13 A). Thus to block tectal responses, the injection electrode had to be placed at different sites in the ICC in device-reared versus normal owls (Fig. 13 B).

Fig. 13.

Summary of the effect of lidocaine injection in the lateral shell of the ICC on auditory responses in the optic tectum of normal (gray symbols) and device-reared (black symbols) owls. A: responses of tectal units to a narrowband stimulus, set at the best ITD and best ILD of the ICC injection site, after 15–90 min of lidocaine injection are plotted as a function of responses to the same stimulus without lidocaine injection (averaged from baseline and recovery). Circles, data obtained using a ∼4-kHz stimulus that matched the frequency and ITD tuning at the ICC injection site. Triangles, data obtained using a ∼8-kHz stimulus that did not match the frequency tuning of the ICC's injection site. For these data, ITD tuning at the ICC injection sites matched, within 14 μs, the abnormal, ITD tuning at the tectal recording site. Squares, data from tectal sites that were tuned to a 4-kHz ITD that did not match the best ITD at the ICC injection site. Symbols and error bars represent means ± SE (n = 20–100 repetitions). Symbols to the left or right of the diagonal dashed line indicate that responses during drug application were greater than or less than, respectively, responses otherwise. Filled symbols, responses during drug application were significantly different from both baseline and recovery responses (unpaired t-test, P < 0.01) but that baseline and recovery were not different from each other.B: for the “matched” data in A, the visual RF azimuth of the tectal recording site is plotted as a function of the best ITD at the ICC injection site. Filled symbols, sites with significant effects, as in A. Note that the recording sites were each located in the rostal portion of the tectum, but the location of the matched injection sites in the ICC were substantially further caudal in device-reared vs. normal owls. The dashed line indicates the predicted relationship in normal owls (Brainard and Knudsen 1993).

DISCUSSION

This study demonstrates that abnormal auditory experience can cause plasticity in the functional connectivity between neurons in the barn owl's inferior colliculus. This plasticity can account for shaping the midbrain map of auditory space to reflect the experience of the individual. In the sections that follow, we first review the pattern of connectivity in the inferior colliculus and optic tectum of normal owls, focusing on how this pattern contributes to the representation of ITD in these nuclei. We then discuss the present results that indicate experience-induced plasticity in this pattern of connectivity. Finally, we consider the implications of this type of plasticity for the connectivity within the inferior colliculus.

Connectivity in normal owls

In normal owls, the representation of ITD in the inferior colliculus and optic tectum is maintained along systematic connections within and between these nuclei (see Figs. 1 and 2 A). The inferior colliculus receives timing information from the nucleus laminaris, where interaural phase differences are first measured (Carr and Konishi 1988, 1990; Sullivan and Konishi 1986). This information is received by the core subdivision of the ICC (Takahashi and Konishi 1988), which projects, in turn, to the contralateral ICC lateral shell (Takahashi et al. 1989). ITD information from the lateral shell converges across frequency channels in the projection from the ICC to the ICX. Thus ICX neurons are broadly tuned for frequency and are tuned for ITD. These neurons are organized with respect to their tuning to sound localization cues, including ITD and ILD, to form a physiological map of auditory space (Mogdans and Knudsen 1993; Moiseff and Konishi 1981). The ICX space map, including the systematic ITD representation along its rostrocaudal dimension, is conveyed via point-to-point projections to the optic tectum (Knudsen and Knudsen 1983).

This pattern of connectivity in the inferior colliculus and optic tectum predicts that the representation of a certain frequency-specific ITD is conveyed from a particular site in the lateral shell to a site in the ICX and, from there, to a site in the tectum. Accordingly, the present results demonstrate that, in a normal owl, elimination of neuronal activity at a critical site in the ICC that is tuned for a certain combination of frequency and ITD causes the selective elimination of responses in the tectum to that same combination of frequency and ITD (Figs. 5-7). This result supports the idea that the pattern of ICC-ICX-tectum connectivity is responsible for the representation of ITD in the ICX and tectum.

Experience-induced changes in connectivity

In device-reared owls, as in normal owls, elimination of neuronal activity at a critical site in the ICC tuned to a certain combination of frequency and ITD blocked auditory responses of neurons in the tectum only for that frequency and ITD (Figs. 8-13). However, the critical site of injection in the ICC was at vastly different locations in device-reared versus normal owls (Fig. 13 B) because device rearing dramatically altered the representation of ITD in the tectum but not in the ICC (Gold and Knudsen 2000a,b). These results indicate that adaptive adjustment to the device corresponded to a change in the pattern of connectivity between the ICC and tectum.

In principle, this plasticity could occur at any point along the pathway from the ICC to the tectum, which are not directly connected (see Fig. 1). Nevertheless, several lines of evidence suggest that the plasticity occurs in the ICC-ICX projection. The first line of evidence is that device rearing does not affect the representation of ITD in the ICC but causes similar changes in the two succeeding nuclei in the ascending pathway, the ICX and optic tectum (Gold and Knudsen 2000a,b). A straightforward explanation is that plasticity in the ICC-ICX projection causes the abnormal representation of ITD in the ICX, which is conveyed directly, along unmodified connections, to the tectum.

The second line of evidence is based on the short latencies of adaptive responses in the ICX. We reported previously that device-induced changes in the representation of ITD in the ICX are evident in the earliest responses of ICX units (Gold and Knudsen 2000b). This result demonstrates that the expression of functional changes in the ICX do not require signals from neurons in other, more remote parts of the CNS (e.g., the cerebellum or forebrain) (Cohen and Knudsen 1999) that would take longer to arrive, implying that this adaptive plasticity involves changes along the ICC-ICX pathway.

The third line of evidence for plasticity in the ICC-ICX projection is from experiments in owls raised with abnormal vision. Like device rearing, raising owls with prismatic spectacles that optically shift the visual field results in a systematic shift in the representation of ITD in the ICX and optic tectum but not in the ICC (Brainard and Knudsen 1993). Recent experiments show that these functional changes are associated with plasticity occurring within the inferior colliculus, including the formation of novel anatomical ICC-ICX connections (DeBello et al. 2001; Feldman and Knudsen 1997), changes in the strength of existing connections (Knudsen 1998), and changes in the balance of excitation and inhibition (Zheng and Knudsen 1999). Considering that these visually driven mechanisms normally function to calibrate auditory spatial information, all of these mechanisms are ideal candidates for driving the adaptive, functional changes observed here in device-reared owls.

Implications for connectivity in the inferior colliculus

The present study focused on changes in the representation of 4-kHz ITDs in the right inferior colliculus and optic tectum because those representations are well characterized in both normal and device-reared owls. In principle, other device-induced changes in auditory tuning could result from a similar form of plasticity. For example, device rearing causes a substantial reduction in normally robust responses to stimuli near 6 kHz in the ICX and tectum. This reduction in responses to 6-kHz stimuli, like the loss of responses to normal ITD values in owls raised with abnormal hearing or vision (Brainard and Knudsen 1993; Gold and Knudsen 2000a,b), could reflect a loss of functional connections from that frequency channel in the ICC. In addition, device rearing shifts the ICX and tectal representations of ITD at 8 kHz by ∼15 μs, in the direction opposite to the shift in the ITD representations at 4 kHz (Gold and Knudsen 2000a,b). A corresponding change in the representation of ITD at 8 kHz does not occur in the ICC and therefore probably reflects novel connections from the ICC to the ICX. These different effects for different frequencies imply that the connectivity within this pathway can be shaped extensively on a frequency-by-frequency basis.

Acknowledgments

We thank M. Shadlen for helpful comments on the manuscript and P. Knudsen for expert technical assistance.

This work was supported by National Institute on Deafness and Other Communication Disorders Grant R01 DC-00155-18.

Footnotes

  • Present address and address for reprint requests: J. I. Gold, Dept. of Physiology and Biophysics, University of Washington Medical School, Box 357290, Seattle, WA 98195-7290 (E-mail:jig{at}u.washington.edu).

REFERENCES

View Abstract