Abstract

Neurons of similar frequency preference are arranged in isofrequency bands (IFBs) across the primary auditory cortex (AI) of many mammals. Across the AI of the cat, one of the most frequently studied species for auditory anatomy and function, we demonstrate IFB-like responses using optical imaging of intrinsic signals (OIS). Optically defined activations were extensively elongated along the dorsoventral axis of AI (the ratio of the major and minor axes was ∼2:1), and systematically shifted as a function of stimulus frequency. The elongation of this IFB-like zone was more conspicuous at higher frequencies. In the ventral sector of the imaged field, the IFB-like zones of activation evoked at different pure tone frequencies tended to overlap extensively. Electrophysiological recording from loci within the optically defined zones of activation revealed matched responses to the frequencies used for optical imaging at 65% of these loci. The dorsoventral orientation of these zones of activation was also closely matched with the orientation of tangentially spreading intrinsic axon terminals, as revealed anatomically. The visualization of IFB-like architecture and tonotopic organization by OIS provides a basic framework for investigating the relationships of different spectral channels and between multiple acoustic parameters at a neuronal population level.

Introduction

In many mammalian species, including humans, primary auditory cortex (AI) is characterized by systematic shifts of frequency representation, tonotopy (Merzenich and Brugge, 1973; Merzenich et al., 1973, 1976; Hellweg et al., 1977; Imig and Adrian 1977; Reale and Imig, 1980; McMullen and Glaser, 1982; Aitkin et al., 1986; Kelly et al., 1986; Pantev et al., 1988; Sally and Kelly, 1988; Morel et al., 1993; Thomas et al., 1993; Howard et al., 1996; Kosaki et al., 1997; Lütkenhöner and Steinsträter, 1998; Formisano et al., 2003; Talavage et al., 2004). Other acoustic parameters, such as the intensity threshold, dynamic range, best sound pressure level associated with intensity function and binaural interactions, have been suggested to be represented along an axis orthogonal to the isofrequency (IF) dimension (Imig and Adrian, 1977; Middlebrooks et al., 1980; Heil et al., 1992, 1994; Schreiner and Sutter, 1992; Schreiner et al., 1992; Clarey et al., 1994; Nakamoto et al., 2004).

An array of spectrotemporal filters, corresponding to a spatial gradient of dominant frequencies across AI, has been demonstrated by several measures. These include local potential changes of a population of activated neurons (field potential; Woolsey and Walzl, 1942), action potentials recorded from single or clustered neurons (single- or multiple-unit recording; Merzenich and Brugge, 1973; Merzenich et al., 1973), activity-dependent accumulation of metabolites (2-deoxyglucose method; Caird et al., 1991), membrane potential indicators (voltage-sensitive dye imaging; Taniguchi et al., 1992; Uno et al., 1993) and intrinsic signals related to hemodynamic changes (Bakin et al., 1996; Hess and Scheich, 1996; Dinse et al., 1997; Harrison et al., 1998; Harel et al., 2000; Spitzer et al., 2001; Tsytsarev and Tanaka, 2002; Versnel et al., 2002; Nelken et al., 2004).

The optical imaging of intrinsic signals (OIS) measures wavelength-dependent changes in hemoglobin concentration, light scattering, cell swelling and other physiological events (Cohen, 1973; Frostig et al., 1990). An illumination light at a wavelength range of 540–570 nm detects blood volume changes, while an illumination light at a wavelength range of 600–630 nm detects a balance of deoxyhemoglobin and oxyhemoglobin increases coupled to transient changes of blood flow (Frostig et al., 1990; Sheth et al., 2004). Functional architecture in the visual and somatosensory cortices has been successfully visualized using OIS at these wavelengths.

Optical demonstration of frequency representations comparable to that defined electrophysiologically will provide a basis for understanding functional architecture in the auditory cortex. Such a fundamental architecture could be correlated to other functional organizations represented spatially (see Schreiner et al., 2000; Read et al., 2002). Optically defined activation could also be related to anatomical connections that underlie the processing of different spectral and temporal parameters (Matsubara and Phillips, 1988; Ojima et al., 1991; Wallace et al., 1991; Read et al., 2001; Ojima and Takayanagi, 2004).

Optical signals acquired from auditory cortex, however, are considered different from the signals evoked from other sensory cortices. For example, OIS in carnivores, including the cat and ferret (Dinse et al., 1997; Spitzer et al., 2001; Versnel et al., 2002; Nelken et al., 2004), was successful only at a 540 nm. Optical imaging from rodents, including the rat, guinea pig and gerbil (Hess and Scheich, 1996; Bakin et al., 1996; Schulze et al., 2002), utilized a 630 nm illumination light. Imaging from the rodent may also be possible at 540 nm (chinchilla; Harrison et al., 1998; Harel et al., 2000). Furthermore, although OIS has demonstrated a frequency-dependent shift of activation, the activation cannot be regarded as equivalent to frequency bands defined electrophysiologically following pure tone (PT) stimulation (i.e. the isofrequency band; IFB). Almost all imaging studies, including those in cats (Spitzer et al., 2001), have displayed the activation areas as circular or oval in shape. This contrasts to the elongated configuration of electrophysiologically determined IFBs (Merzenich et al., 1973).

These previous studies might have suggested certain limitations in applying the technique to a characterization of the functional architecture of auditory cortex. However, recent studies which have reported linear relationships between the intensity of optical signals and the magnitude of neuronal activities (Sheth et al., 2003, 2004; Nemoto et al., 2004) drove us to reconsider the means for evoking larger neuronal activities from the auditory cortex. Larger neuronal activities are likely to induce a greater strength of optical signals, and this would lead to well-localized areas of activation after averaging repeated trials, comparable to the organizations that have been defined electrophysiologically. Here, by stabilizing the animal conditions and optimizing stimulus and recording procedures, we have visualized IFB-like zones of activation that are dorsoventrally elongated following stimulation with trains of PT pips. Superimposition of zones of activation evoked at different frequencies reveals that they overlap significantly in the ventral one-third of the entire activation field. Furthermore, activation pattern revealed by the optical imaging were compared with electrophysiologically recorded multi-unit responses. Finally, tracer injections were performed to compare the tangential distribution of intrinsic cortical connections and the orientation of the elongated zones of activation which were revealed optically. The dorsoventral elongated zones of activation are almost exactly aligned with the overall distribution of the labeled axon terminals.

Materials and Methods

Animals and Surgery

Experimental procedures for OIS consisted of three steps: first, putting the chamber on the skull; second, capturing optical images in multiple sessions; and third, off-line data analysis. In some cases, the experiments included tracer injections or electrophysiological mapping. The experimental protocol was approved by the Experimental Animal Committee of the RIKEN Institute, and was in accordance with the National Institutes of Health (USA) Guide for the Care and Use of Laboratory Animals.

All surgical and imaging procedures were carried out under aseptic conditions. Cats (n = 10, both sexes, weighing 1.7–4.0 kg) were administered atropine sulfate (0.15 mg/kg i.m.) and anesthetized with sodium pentobarbital (Nembutal, Abbott, TX; 35–40 mg/kg i.p.). Body temperature was maintained by a thermistor-controlled heating pad at 37–38°C. A supplemental dosage of sodium pentobarbital was infused through an intravenous catheter (2–5 mg/kg/h i.v.). Animal state was monitored by electroencephalogram (EEG, band-passed at 0.5–100 Hz), heart rate, breathing rate, pupil size and rectal temperature, together with frequent confirmation of the absence of response to strong forepaw pinching. A small metal anchor was glued to the dorsal skull of the animal with dental resin. Screws were implanted for epidural EEG recordings. The animal was fixed to a post via the head anchor, and a round chamber (20 mm in diameter, 4 mm in height) was attached to the lateral skull with dental resin. At the end of the surgery, the skin was vigorously rinsed with saline, followed by local application of an antibiotic (Kenicef, 40 mg/kg i.m.) and lidocaine (Xylocaine). The antibiotic and analgesic (Menamin, 1–2 mg/kg i.m.) were injected systemically as needed. Animals were maintained under warm conditions until they were able to begin to walk.

For optical imaging, typically 1–3 weeks after the chamber mounting (the first imaging experiment), cats were initially medicated with atropine sulfate (0.15 mg/kg i.m.) and droperidole (Droleptan, 0.4 mg/ml i.m.) followed by ketamine hydrochloride (Ketalar, 16–20 mg/kg i.m.). After a tracheal intubation, the animal was fixed to a post via the previously implanted head anchor, intravenously catheterized, muscle-relaxed with pancuronium bromide (Mioblock, ∼0.04 mg/kg/h i.v.) dissolved in glucose-electrolytic solution (Soldem 3AG) supplemented with riboflavin (Bislase, 0.2 mg/kg/h), and artificially ventilated with a mixture of N2O and O2 (2:1 in volume) and isoflurane (Forane). The skull inside the chamber was sterilized and removed, dura mater was carefully dissected, and agarose HGS (1.25% in saline, 36°C) containing dexamethazone (0.04 mg/ml) and antibiotics (Lincocin, 4 mg/ml) was quickly introduced into the chamber. Finally, the craniotomy was sealed with a glass cover.

The anesthetic level was continuously monitored via the EEG, heart rate and expired-CO2 level (Capnomac Ultima, Datex-Ergstrom, Finland). During the craniotomy, anesthetic depth was adjusted so that EEG pattern retained slow waves (<10 Hz) of large amplitudes. Typically, the isoflurane concentration was between 1.0 and 2.0%. During imaging following the craniotomy, the anesthetic level was carefully adjusted so as to keep slow waves superimposed on the higher frequency (>10 Hz) waves of smaller amplitude (Ikeda and Wright, 1974; Schwender et al., 1998; Villeneuve and Casanova, 2003). The isoflurane concentration was typically between 0.6 and 1.5%. In addition, at every 30–50 min, forepaw pinching and eyelid touching was performed to see if it quickly affected the animal's heart rate (Villeneuve and Casanova, 2003) and/or EEG pattern.

Optical Imaging

Imaging was carried out in a sound-attenuated room (−45 dB at 500 Hz). All instruments other than a CCD camera (CS8310, Tokyo Electronic Industry, Japan) and light guides were placed outside the room. Intrinsic signals were recorded through the glass cover at the top of the chamber. Images were enlarged through a face-to-face tandem combination of two 50 mm lenses (Nikon, F = 1.2; Ratzlaff and Grinvald, 1991) attached to the camera and were digitized with an IBM/PC-compatible computer equipped with a video frame grabber board (Pulsar, Matrox, Quebec, Canada). The imaged field had a dimension of 9.0 × 6.7 mm (320 × 240 pixel). The surface blood vessel pattern was captured by illumination light at a wavelength (mean ± SD) of 540 ± 10 nm before functional imaging. The focal plane of the camera was then lowered to ∼600 μm below the cortical surface to minimize artifacts due to the presence of surface blood vessels. Optical signals induced by acoustic stimulation were captured with a 540 ± 10 nm illumination light for all cases. In two hemispheres, image capturing was attempted also at 610 ± 10 nm, but no cortical activation was detected (Spitzer et al., 2001; Nelken et al., 2004). Therefore, images shown in this study were all captured at 540 nm.

One session consisted of 30 blocks, with each block comprised five trials at different frequencies (0 = silence, 21, 4, 14 and 8 kHz; Fig. 1). These different frequency values were presented in a circular order (as above) in most sessions or in a randomized order in a few sessions. No difference was found under these two conditions. A single trial consisting of a sequence of 16 image frames (each lasting 0.5 s in duration) was repeated once per 21 or 23 s. The tone stimulation was started synchronously with the onset of the third image frame. The sound intensity level was fixed throughout the session.

Figure 1.

Block diagram showing the temporal sequence of illumination light on/off, image capturing, and stimulus on/off. Optical imaging at five different frequencies (f1–f5, including one for silence) per block is repeated 30 times. An averaged image of frames 1–2 is used as a reference image. Frames 3–11 (data frames) are averaged to acquire a single intensity map (see Fig. 2ad) for each stimulus frequency. The interval of the stimulation is either 21 or 23 s.

Figure 1.

Block diagram showing the temporal sequence of illumination light on/off, image capturing, and stimulus on/off. Optical imaging at five different frequencies (f1–f5, including one for silence) per block is repeated 30 times. An averaged image of frames 1–2 is used as a reference image. Frames 3–11 (data frames) are averaged to acquire a single intensity map (see Fig. 2ad) for each stimulus frequency. The interval of the stimulation is either 21 or 23 s.

Acoustic Stimulation

Acoustic stimuli were applied binaurally through short silicon tubes connecting speakers to the external ear canals. A tight seal was secured by putting Vaseline between the silicon tube and ear canal. Pure tone pips were generated digitally by a MALab system (Kaiser Instruments, CA) controlled by a Macintosh computer (Krishna and Semple, 2000; Ojima and Murakami, 2002). The sound delivery system coupled to a custom-made dummy ear canal was calibrated using a pressure-type condenser microphone (7017, ¼ inch, ACO, Japan), with reference to a standard sound pressure level (SPL; 94 dB re 20 μPa, at 1 kHz). The calibration data for each ear were stored in a computer file for use in controlling a digital attenuator to obtain the desired SPLs.

A single stimulus was a train of 10 repeats of PT pips (50 ms duration, 10 ms rise/fall time, and an interval of 0.2 s, lasting for 2.0 s). In most cases, 60 dB SPL was used for evoking optical responses, but various SPLs ranging from 30 to 70 dB were also applied at a 10 dB step in one session for examining the intensity-dependency of optical responses.

Electrophysiology

Extracellular multiunit recording of responses to PT pips was carried out (three hemispheres) in the second imaging experiment, performed 3–4 days after the first imaging experiment. At the end of the first experiment, the outer surface of the head-attached chamber (with the agarose kept inside) was rinsed vigorously with saline containing Iodine. The chamber was capped with a plastic lid. The animals were administered antibiotics (Kenicef, 40 mg/kg i.m.) and analgesia (Menamin, 1–2 mg/kg i.m.). Following disconnection of the intravenous catheter, they were allowed to recover from the anesthesia and were returned to a house cage, with free access to food and water. For the following 3–4 days, antibiotic and analgesia were administered as needed. The basic procedures for electrophysiological determination of the frequency response ranges were described previously (Ojima and Takayanagi, 2004). Briefly, following several imaging sessions, a reference image of the cortical vasculature pattern was used to guide electrode penetrations and for the later alignment of the optical and electrophysiological data. The glass cover on the chamber top was replaced by a silicon rubber ring with a large opening. Heavy silicon oil (5k cs, Shinetsu Chemical, Japan) was put on the agar surface to prevent it from drying, and a tungsten microelectrode (2–3 MΩ impedance at 1 kHz, FHC, ME), held on a stepping motor microdrive (PC-5N, Narishige, Japan), was introduced into the cortex perpendicular to its surface. The recording depth was 600–800 μm below the cortical surface, as judged by digital read-out of the microdrive. Signals were filtered (band-passed from 0.5 to 3.0 kHz) and fed to a window discriminator to make raster display. At each recording point, a frequency response range was automatically determined (MALab) using PT pips (50 ms duration, 5 ms rise/fall time and 10 repeats at 0.9 Hz) at the SPL used for optical recording and also at SPLs 10–20 dB above the minimum threshold for many cases. Spikes were counted from 10 to 100 ms after stimulus onset for construction of peristimulus time histograms.

Tracer Injections and Histological Processing

The basic procedures for tracer injection and histochemical/immunohistochemical visualization of tracers were described in detail previously (Ojima and Takayanagi, 2001). Unilateral tracer injections were made following either optical imaging or electrophysiological mapping. Briefly, after removing the glass cover on the chamber, injections of two anterograde tracers, biotinylated dextran amine (BDA; Molecular Probes, OR) and Phaseolus vulgaris leucoagglutinin (PHA-L; Vector, CA) were made iontophoretically (positive 6 μA with a 7 s on/off duty cycle, for 20 min). The injections were targeted at loci within the activation zones that had been defined previously in the optical recordings and centered at a depth of ∼700 μm from the cortical surface. One tracer was placed in a lower frequency activation zone, while the other was placed in a higher frequency zone. For the injection, glass micropipettes (A-M Systems, MA) with an outer tip diameter of 20 μm were used. For more precise alignment of the cortical surface image and anatomical labeling in tangential sections, a DiI (Molecular Probes, OR)-coated tungsten microelectrode was inserted perpendicularly to the cortical surface at three points. This left a fluorescent tracer that could be used in aligning data sets. The exposed cortex was rinsed, covered with a patch of dura substitute (GORE-TEX PATCH, Gore & Associates, AZ) and closed with dental resin. Animals were then cared according to the protocol of post-surgical treatment as described above.

After electrophysiological recording or 2–3 weeks following tracer injections, animals were deeply anesthetized with sodium pentobarbital (Nembutal, 60 mg/kg i.P.) and were perfused transcardially with 300 ml saline followed by 2 liters of chilled fixative (4% paraformaldehyde in phosphate buffer, pH 7.4). The brain was removed and cut in serial tangential sections (50 μm in thickness). Alternative sections were processed for either BDA or PHA-L; BDA was reacted with ABC reagent (Vector Laboratories, OR) and PHA-L in the steps of PAP method. Both tracers were visualized by diaminobenzidine (Sigma, MO) as brown color reaction products. Labeled axon terminals (varicosities) were plotted as dots across a tangential section passing through layer 3 (∼ 600–700 μm in depth; Winer, 1984) using Neurolucida (MicroBrightfield, VT). The plot images were aligned with optical images in relation to the three reference points of DiI observed under a fluorescent microscope.

Data Analysis

Off-line analysis of image frames was carried out with a commercial software package IDL (Research Systems Inc., CO). Intrinsic signals (darkened areas, corresponding to decreased changes in light reflectance) were visualized as follows. First-frame analysis was carried out to remove slowly changing vascular noise from captured images; each frame obtained after stimulus onset was divided by an average of the prestimulus frames (frame nos 1 and 2; Shoham and Grinvald, 2001). Typically, the third to eleventh frames (data frames) were averaged for the spatial extent of activated areas (intensity map, see Fig. 2a–e). Only a high-cut Gaussian spatial filter (cut-off frequency σ = 14 cycle/mm) was used to eliminate high frequency noise. One intensity map was acquired for each stimulus frequency by averaging 30 images stored in one session (e.g. Fig. 2a–e).

Figure 2.

Intensity maps (a–e), statistical maps (f–i, k–n, p–s, and u–x) and composite maps (j, o, t and y) obtained from the cat middle ectosylvian gyrus by using optical imaging of intrinsic signals. The illumination light wavelength is 540 nm. Trains of pure tone (PT) pips evoke decreases in light reflectance (darkened areas in a–d; see e for scale) in restricted areas. Panels f–i (4, 8, 14 and 21 kHz in this order) are statistical maps defined by comparing the intensity maps acquired under the PT stimulus condition (a–d) with that under the non-stimulus condition (e). Pixels activated significantly (t-test, P < 0.01) by stimulation are colored: differential colors (purple, yellow, green and red) are used for the four PT frequencies (4, 8, 14 and 21 kHz, respectively; f–i). The zones of activation evoked at these frequencies are superimposed onto a single, black-and-white surface view to display the tonotopic arrangement (j). Panels k–o are statistical maps obtained from a session repeated 15 min after the preceding session (f–j). Two other hemispheres are illustrated in the fourth (p–t) and fifth (u–y) columns according to the same format. Horizontal and vertical arrows in e, t and y point anteriorly (a) and dorsally (d). aes, anterior ectosylvian sulcus; pes, posterior ectosylvian sulcus. The color representation of the different stimulus frequencies, the directions of arrows and the abbreviations of sulci apply to all other figures. Imaged areas: 6.7 mm (vertical) × 9.0 mm (horizontal). The suprasylvian sulcus is oriented horizontally.

Figure 2.

Intensity maps (a–e), statistical maps (f–i, k–n, p–s, and u–x) and composite maps (j, o, t and y) obtained from the cat middle ectosylvian gyrus by using optical imaging of intrinsic signals. The illumination light wavelength is 540 nm. Trains of pure tone (PT) pips evoke decreases in light reflectance (darkened areas in a–d; see e for scale) in restricted areas. Panels f–i (4, 8, 14 and 21 kHz in this order) are statistical maps defined by comparing the intensity maps acquired under the PT stimulus condition (a–d) with that under the non-stimulus condition (e). Pixels activated significantly (t-test, P < 0.01) by stimulation are colored: differential colors (purple, yellow, green and red) are used for the four PT frequencies (4, 8, 14 and 21 kHz, respectively; f–i). The zones of activation evoked at these frequencies are superimposed onto a single, black-and-white surface view to display the tonotopic arrangement (j). Panels k–o are statistical maps obtained from a session repeated 15 min after the preceding session (f–j). Two other hemispheres are illustrated in the fourth (p–t) and fifth (u–y) columns according to the same format. Horizontal and vertical arrows in e, t and y point anteriorly (a) and dorsally (d). aes, anterior ectosylvian sulcus; pes, posterior ectosylvian sulcus. The color representation of the different stimulus frequencies, the directions of arrows and the abbreviations of sulci apply to all other figures. Imaged areas: 6.7 mm (vertical) × 9.0 mm (horizontal). The suprasylvian sulcus is oriented horizontally.

Activation areas were statistically defined for each stimulus condition (4, 8, 14 or 21 kHz PT). The statistical comparison was made using Student's t-test; the mean signal intensity and SD of each pixel across all blocks (n = 30) in one session were calculated, and compared pixel-by-pixel between the non-stimulus and stimulus conditions typically at a significance level of P < 0.01. Regions overlapped by different frequency zones were defined by superimposing zones of activation determined at a significance level of P < 0.0001 or P < 0.00001. Thereafter, for simplicity, binning (4 × 4 pixels averaged and reduced to a single pixel) was carried out, and isolated clusters consisting of <5 activated pixels were deleted from the statistically defined maps. Zones of activation evoked by the stimulus set of four different frequencies were colored differentially (e.g. Fig. 2f–j) and then superimposed on a black-and-white cortical surface view taken before the session.

The major and minor axes of activation zones were determined using the public domain NIH Image (National Institutes of Health) according to a strategy of the best fitting ellipse. Angle of the major axis of ellipse was also defined and was regarded as the orientation of the zone of activation.

For the overall orientation of distribution of labeled axon terminals, each dot representing a single axon terminal in the anatomical plotting (Fig. 7) was enlarged until it reached the diameter equivalent to 16 pixels. This resulted in the fusion of the enlarged dots that were located in close proximity, and generated one or two large islands of the fused dots as well as many small islands. In each injection, the best fitting ellipse analysis was applied to the largest island. The angle of its major axis relative to the x (horizontal) axis of the imaged field was defined, and was regarded as the overall orientation of the extent of the axon terminal distribution.

Results

In eight out of 10 hemispheres (five left and three right), cortical activation (decrease in light reflectance) following PT stimulation and its topographical shift were repeatedly detected in successive and/or different sessions performed on the same or different day. All stimulus frequencies were effective at least at 60 dB SPL in evoking optical responses. The extent and contour of the activation, however, varied to a certain degrees from session to session. In the remaining two hemispheres, no cortical activation or a non-reproducible activation pattern was induced, giving an impression that the brain appeared to have been damaged during the surgical preparation for OIS. These cases were excluded from the data analysis.

Areas of Activation Following Stimuli with Trains of PTs of Different Frequencies

Images of cortical activation following PT stimulation at foru different frequencies are shown in three hemispheres (Fig. 2a–o,p–t,u–y). In the intensity maps (Fig. 2a–e), areas showing reduced light reflectance (i.e. activation) appear as darkened zones. The darkened zones were conspicuously elongated approximately dorsoventrally or slightly tilted from the dorsoventral axis. Under the non-stimulus (silent) condition (Fig. 2e), the cortical surface appears to be almost homogeneously gray across the imaged field, with only occasional spots of substantially increased darkening beyond the background. These areas correspond to thick blood vessels coursing perpendicular to the cortical surface.

Statistical comparison of these intensity maps obtained under both stimulus and silent conditions more clearly points out the extent of the activated zones. Differentially colored zones of activation (Fig. 2f–i,k–n,p–s,u–x), representing the optically defined responses to stimulation at fourdifferent PT frequencies, were superimposed onto a single cortical surface view to facilitate tonotopic mapping (Fig. 2j,o,t,y). As evident also in the intensity maps (Fig. 2a–d), the statistical maps showed zones of activation elongated along the dorsoventral axis or an axis slightly tilted from this, with a tendency for higher stimulus frequencies to evoke more noticeably elongated zones. The averaged major/minor axes of the zones of activation within the imaged field (mean ± SD in mm; n = 8) were 3.63 ± 1.18/1.98 ± 0.63, 4.39 ± 1.06/2.02 ± 0.86, 5.28 ± 1.62/2.71 ± 0.44 and 5.94 ± 1.49/2.55 ± 0.66, in increasing order of frequency of the stimulus set (i.e. 4, 8, 14 and 21 kHz). Zones of activation occasionally had a dumbbell-like configuration along the major axis (Fig. 2g,n,q,r,v) or split into two compartments, one located more dorsally and the other more ventrally (Fig. 2l,x).

The third column (Fig. 2k–o) shows a repeated session conducted 15 min after the end of the preceding session shown in the second column (Fig. 2f–j). Comparison between the second and third columns indicates a good, but not perfect, correspondence in terms of the extent and contour of activation obtained in different sessions (also see Fig. 5c,e).

Overlap of Zones of Activation Visualized at Different Pure Tone Frequencies

As demonstrated in the hemispheres shown in Figure 2j,o,t,y, the adjacent zones of activation induced at different PT frequencies showed partial overlap. The overlap of different frequency zones appeared to be more extensive in the ventral and dorsal sectors (arrowheads) of the entire activation zones than in the central sector. This was confirmed by localizing overlap of activated zones defined at a higher P-value (P = 0.0001 or 0.00001) during the statistical comparison of images obtained under the stimulated and non-stimulated conditions. As shown in Figure 3A, it was revealed that regions of overlap (black) were confined to the ventral sector of activated zones (and, to a lesser extent, in the dorsal sector), but not in the central sector in most hemispheres. The absence of overlap in the central sector could also be illustrated by projecting the overlap and non-overlap regions onto the dorsoventral axis of the imaged field (see an example on the right of panel 8 in Fig. 3A; black indicates overlap and gray non-overlap). There was an extensive non-overlap partition (arrowheads) in the center of the vertical scale, although its extent varied from hemisphere to hemisphere, as shown in Figure 3B. Furthermore, when the entire activation field was divided into three sectors of an equal dorsoventral dimension, more of the overlapped areas were localized in the ventral sector compared to the central sector (P < 0.0005, Welch's test; Fig. 3C).

Figure 3.

Regions overlapped by zones of activation induced at different stimulus frequencies. (A) Zones of activation defined statistically (see Materials and Methods) are outlined by solid lines, and the overlapped regions are depicted by filled pixels (black). The suprasylvian sulcus is oriented horizontally. Eight hemispheres (1–8) are shown. (B) In each hemisphere, these overlapped regions are projected onto the vertical scale equivalent to the dorsoventral dimension of the entire activation field (see the example in A8). An extensive non-overlap segment (the gray part indicated by arrowheads) tends to be located in the center of the scale in each case. (C) The tendency for the overlap to be made ventrally and/or dorsally between the zones activated at different frequencies is analyzed statistically. The entire activation field is divided into dorsal, central and ventral sectors of an equal vertical dimension (as exemplified by the dotted line in panel A2). The mean of the overlapped pixels calculated for each sector across all eight cases is compared statistically (Welch's test). Note that the ventral zone has more overlapped pixels than the central zone (P < 0.0008).

Figure 3.

Regions overlapped by zones of activation induced at different stimulus frequencies. (A) Zones of activation defined statistically (see Materials and Methods) are outlined by solid lines, and the overlapped regions are depicted by filled pixels (black). The suprasylvian sulcus is oriented horizontally. Eight hemispheres (1–8) are shown. (B) In each hemisphere, these overlapped regions are projected onto the vertical scale equivalent to the dorsoventral dimension of the entire activation field (see the example in A8). An extensive non-overlap segment (the gray part indicated by arrowheads) tends to be located in the center of the scale in each case. (C) The tendency for the overlap to be made ventrally and/or dorsally between the zones activated at different frequencies is analyzed statistically. The entire activation field is divided into dorsal, central and ventral sectors of an equal vertical dimension (as exemplified by the dotted line in panel A2). The mean of the overlapped pixels calculated for each sector across all eight cases is compared statistically (Welch's test). Note that the ventral zone has more overlapped pixels than the central zone (P < 0.0008).

Time Course of Activation

The spatio-temporal development of activation was analyzed sequentially in 1.0 s time window. Figure 4A is a set of image frames from a single session and activation is initiated in the fourth/fifth frame (i.e. second/third data frame after stimulus onset) in response to 4, 8 and 21 kHz stimulation, or in the fifth/sixth frame (i.e. third/fourth data frame) in response to 14 kHz stimulation. In the frames prior to these, no activation was present, and appears as a background image in a 0.5 s time window (i.e. third frame for 4, 8 and 21 kHz and third and fourth frames for 14 kHz). On average, the activation was initiated in frame no. 2.2 (SD = ±0.5, n = 32).

Figure 4.

Temporal sequence of activation evoked by PT stimulation. (A) A set of activations arranged in consecutive frames (horizontal row) from the time of stimulus onset (third frame) to the end of image capturing (sixteenth frame) is shown for each stimulus frequency. The numeral(s) stands for the frame number(s). Before the onset of activation, the background cortical surface view(s) is shown (e.g. third frame for 4 kHz, and third and fourth frames for 14 kHz). Note the activation initiated at two segregated loci located dorsally and ventrally (arrowheads) in the onset frame (e.g. the fourth/fifth frame for 4 kHz and the fifth/sixth frame for 14 kHz). (B) Temporal sequence of reflectance change in the pixel showing peak activation in the statistical map (–ΔR/R%). The horizontal bar in each panel indicates the stimulus period (between the third to sixth frames). (C) Frames containing the onset activation for 4 different stimulus frequencies are shown in one vertical row. Eight cases are shown. Note that the activation is initiated at two (or more) loci for the stimulation at higher frequencies (8, 14 and 21 kHz), but not for the stimulation at the lowest frequency (4 kHz).

Figure 4.

Temporal sequence of activation evoked by PT stimulation. (A) A set of activations arranged in consecutive frames (horizontal row) from the time of stimulus onset (third frame) to the end of image capturing (sixteenth frame) is shown for each stimulus frequency. The numeral(s) stands for the frame number(s). Before the onset of activation, the background cortical surface view(s) is shown (e.g. third frame for 4 kHz, and third and fourth frames for 14 kHz). Note the activation initiated at two segregated loci located dorsally and ventrally (arrowheads) in the onset frame (e.g. the fourth/fifth frame for 4 kHz and the fifth/sixth frame for 14 kHz). (B) Temporal sequence of reflectance change in the pixel showing peak activation in the statistical map (–ΔR/R%). The horizontal bar in each panel indicates the stimulus period (between the third to sixth frames). (C) Frames containing the onset activation for 4 different stimulus frequencies are shown in one vertical row. Eight cases are shown. Note that the activation is initiated at two (or more) loci for the stimulation at higher frequencies (8, 14 and 21 kHz), but not for the stimulation at the lowest frequency (4 kHz).

Temporal sequences of activation are shown in Figure 4B as a percentage decrease in reflectance of illumination light at the pixel showing the maximum activation (–ΔR/R%) for all eight cases. The activation reached the maximum level in the seventh frame (corresponding to the fifth data frame or a time window of 2.0–2.5 s after stimulus onset). On average, the local activation was maximal in frame no. 5.7 (SD = ±1.7, n = 32). The activation returned to the baseline level before the last frame in some cases (e.g. case 4 in Fig. 4B) or seemed to persist after the last frame in other cases (e.g. case 2 in Fig. 4B).

For the higher frequencies (8, 14 and 21 kHz), the activation tended to be initiated at two or more loci (indicated by arrows in Fig. 4C) in most cases (17 out of 24 stimulus cases). Initiation of the activation at dorsal and ventral loci, however, was observed only in one out of eight stimulus cases for the lowest frequency used (4 kHz). This distinction observed between the 4 kHz stimulation and the 8, 14 and 21 kHz stimulations was statistically significant (P < 0.007, Fisher's exact probability test).

Zones of Activation Evoked at Different Sound Pressure Levels

The zones of activation evoked by the stimulus set at different tonal intensities ranging from 40 to 70 dB SPL are shown in Figure 5a–d, together with one repeat at 60 dB SPL in Figure 5e.

Figure 5.

Effect of sound intensity on the extent and configuration of zones of activation as revealed by optical imaging of intrinsic signals (statistical maps). The overall configuration of activation zones and their systematic shift are similar among four different sound pressure levels (40, 50, 60 and 70 dB SPL in a, b, c and d, respectively), except at threshold and near-threshold pressure levels (40 and 50 dB SPL; see a and b) for the 8 kHz stimulation. At these levels, the stimulation at 8 kHz does not evoke the full extent of activation. An activation image in (e) is a repeated session acquired 15 min after the stimulation with the full set of SPLs. The suprasylvian sulcus is oriented horizontally. Imaged areas: 6.7 mm (vertical) × 9.0 mm (horizontal).

Figure 5.

Effect of sound intensity on the extent and configuration of zones of activation as revealed by optical imaging of intrinsic signals (statistical maps). The overall configuration of activation zones and their systematic shift are similar among four different sound pressure levels (40, 50, 60 and 70 dB SPL in a, b, c and d, respectively), except at threshold and near-threshold pressure levels (40 and 50 dB SPL; see a and b) for the 8 kHz stimulation. At these levels, the stimulation at 8 kHz does not evoke the full extent of activation. An activation image in (e) is a repeated session acquired 15 min after the stimulation with the full set of SPLs. The suprasylvian sulcus is oriented horizontally. Imaged areas: 6.7 mm (vertical) × 9.0 mm (horizontal).

The zones of activation were again typically elongated along the dorsoventrally oriented axis. Their extent and position did not appear to change considerably from one intensity level to another, if the level was above threshold. For example, there were very few or only a slight activation following 8 kHz stimulation at 40–50 dB SPL (Fig. 5a,b), but this turned into an almost fully developed activation by a 10 dB increase in SPL (Fig. 5c). A further increase in SPL up to 70 dB did not result in any dramatic changes in the number or extent of the zones of activation. A repeated session at 60 dB SPL resulted in an activation similar in extent and position to that acquired at the same sound intensity in a preceding session (Fig. 5e).

Electrophysiological Responses and their Correspondence to the Optical Responses

In order to evaluate the correspondence between optically detected activation and the electrophysiological responses, frequency response ranges (FRR) defined by multiunit recordings at the SPL used for optical imaging, were investigated in three hemispheres (Fig. 6).

Figure 6.

Correspondence between optically defined activation zones and the electrophysiological multiunit responses. Three hemispheres are shown (a, b and c). The range of frequency response (numeral) to the SPL used for optical recording (60 dB SPL) is depicted, next to each recording site (dot), on statistical maps obtained at the same SPL used for the optical imaging. Single and double stars in (a–c) indicate examples of the recording points that are located in the single-frequency and two-frequency activation zones, respectively. The spike counts at two representative recording points (indicated by the italicized d and e in b) are shown in (d) and (e), as a peristimulus time histogram (upper) and frequency response curve (lower) at a 60 dB SPL. The stimulus duration is 50 ms (short black horizontal bars in d and e). Spikes are counted from 10 to 100 ms (90 ms in duration) after stimulus onset (the gray boxes in d and e). The suprasylvian sulcus is oriented horizontally. Imaged areas in panels (a) and (b): 6.7 mm (vertical) × 9.0 mm (horizontal). The imaged area in (c) (9.0 × 9.0 mm) is a montage of two statistical maps (upper and lower, divided by broken horizontal lines) taken in successive sessions. Note that the zones activated at different frequency values overlap extensively in the most ventral portion of the entire activation field. NR, no response.

Figure 6.

Correspondence between optically defined activation zones and the electrophysiological multiunit responses. Three hemispheres are shown (a, b and c). The range of frequency response (numeral) to the SPL used for optical recording (60 dB SPL) is depicted, next to each recording site (dot), on statistical maps obtained at the same SPL used for the optical imaging. Single and double stars in (a–c) indicate examples of the recording points that are located in the single-frequency and two-frequency activation zones, respectively. The spike counts at two representative recording points (indicated by the italicized d and e in b) are shown in (d) and (e), as a peristimulus time histogram (upper) and frequency response curve (lower) at a 60 dB SPL. The stimulus duration is 50 ms (short black horizontal bars in d and e). Spikes are counted from 10 to 100 ms (90 ms in duration) after stimulus onset (the gray boxes in d and e). The suprasylvian sulcus is oriented horizontally. Imaged areas in panels (a) and (b): 6.7 mm (vertical) × 9.0 mm (horizontal). The imaged area in (c) (9.0 × 9.0 mm) is a montage of two statistical maps (upper and lower, divided by broken horizontal lines) taken in successive sessions. Note that the zones activated at different frequency values overlap extensively in the most ventral portion of the entire activation field. NR, no response.

Recordings were made at a total of 75 points. Of these 75 recording points, 30 points were in zones activated at single frequencies (e.g. * in Fig. 6a–c). These points revealed a good match between the two recording methods; that is, 80% had a response range that included the frequency value used for the imaging of the activated zones. Another 26 points were in the overlap regions of two or three different zones of activation (e.g. as indicated by ** in Fig. 6a–c). These recording points showed a partial correspondence between the two methods; that is, 38% had a FRR that included only one of the stimulus frequencies used for imaging and another 38% showed a FRR that included two of the stimulus frequencies used for imaging. There were no points that had a FRR which included three frequency values. The remaining 19 points were outside activation zones. By counting the recording points in the two- or three-zone overlap regions in multiple times, we were able to calculate the overall matching rate between the two methods. Approximately 64% of the recording points had a FRR that included the frequency value effective in inducing the optically defined activations.

For a quantitative comparison of the spatial distribution of the FRRs, the FRR index was defined by dividing the higher edge frequency by the lower edge frequency of the range. Multiunit recording points included in the ventral sectors of the entire activation field, on average, had a significantly (Welch test, P < 0.01) larger FRR index (2.42 ± 1.88 octaves, n = 23) than that included in the central sector (1.41 ± 0.34 octaves n = 23). One example of sharply tuned multiple-unit responses located in the central portion of activation zones is shown in Figure 6d. This recording point responded best to a 3.0 kHz PT at the minimum SPL (50 dB, not shown), and showed a narrow FRR of 3.0–3.5 kHz (FRR index, 1.17) at the SPL used for imaging (60 dB SPL). This narrow response range corresponded almost exactly to the frequency value used for the imaging (4 kHz). In Figure 6e, an example of broadly tuned points located in the ventral sector of activated zones is shown. This recording point responded best to a 16.8 kHz PT at the minimum threshold (40 dB SPL), but had a wider FRR of 14.0–20.1 kHz (FRR index = 1.44) at the SPL used for imaging (60 dB SPL).

Distribution of Horizontal Connections

The extent and orientation of the optically recorded activation were compared directly with anatomical intrinsic connectivity by aligning the optical image and the distribution of labeled axon terminals.

Figure 7 shows the overall distribution of labeled axon terminals superimposed on optically defined activation zones (contours in black) in two hemispheres. Two different tracers were injected separately in the center of the 4 kHz zone and in the periphery of the 14 kHz zone of one hemisphere (Fig. 7b,c). The tracers were injected in the center of the 14 and 21 kHz zones of activation of the other hemisphere. The overall distribution of labeled axon terminals (green and purple dots) was, on the whole, oriented roughly along the dorsoventral axis. The angle of the major axis was quantitatively determined for the extent of the axon terminal distribution (see Materials and Methods). It ranged from 71 to 115° (the average of four injections, 86.3°), and was very close to the angle of the vertical (dorsoventral) axis of the imaged field (i.e. 90°).

Figure 7.

Relationships between optically defined zones of activation and the distribution of anatomically labeled axon terminals. Two hemispheres (a–c and d–f) are shown. Intensity map of the first hemisphere (a–c) is shown in (a). For the intensity map of the second hemisphere (d–f), see Figure 6c. Two anterograde tracers are injected in one hemisphere, localized separately in different zones of activation (the circles in red in each hemisphere). Labeled axon terminals (purple and green dots in b–f) originating from two different injection sites are plotted separately on directly adjacent sections passing through middle layer 3, superimposed on the corresponding activation zones (outlined in black) and overlaid on background cortical surface view. Images (d) and (e) are superimposed to illustrate the overlap (black dots indicated by arrowhead) of the differentially colored dots (f). Blue filled circles and yellow crosses are references used to align the anatomical and optical maps. The suprasylvian sulcus is oriented horizontally. Imaged areas: 6.7 mm (vertical) × 9.0 mm (horizontal) for (a–c) and 9.0 × 9.0 mm for (d–f).

Figure 7.

Relationships between optically defined zones of activation and the distribution of anatomically labeled axon terminals. Two hemispheres (a–c and d–f) are shown. Intensity map of the first hemisphere (a–c) is shown in (a). For the intensity map of the second hemisphere (d–f), see Figure 6c. Two anterograde tracers are injected in one hemisphere, localized separately in different zones of activation (the circles in red in each hemisphere). Labeled axon terminals (purple and green dots in b–f) originating from two different injection sites are plotted separately on directly adjacent sections passing through middle layer 3, superimposed on the corresponding activation zones (outlined in black) and overlaid on background cortical surface view. Images (d) and (e) are superimposed to illustrate the overlap (black dots indicated by arrowhead) of the differentially colored dots (f). Blue filled circles and yellow crosses are references used to align the anatomical and optical maps. The suprasylvian sulcus is oriented horizontally. Imaged areas: 6.7 mm (vertical) × 9.0 mm (horizontal) for (a–c) and 9.0 × 9.0 mm for (d–f).

It is apparent that this orientation was almost in parallel with that of the optically recorded activation zone (Fig. 7c–e) or just slightly deviated from it (Fig. 7b). The difference in angle of the extent of the axon terminal distribution and that of the corresponding activation zone was small, and ranged from 7 to 19° (the average of four injections, 13.8°). Along the dorsoventral axis, the elongated extent of the distribution of the labeled axon terminals matched well the elongated extent of the zones of activation in most cases. This was clearly manifested by the distribution of the most dorsal and ventral clusters of labeled axon terminals. They were localized precisely in the very dorsal and ventral portions of the corresponding zone of activation. Consistent with the above observation, a tracer injection in the periphery of an activation zone resulted in a localization of the labeled axon terminals only in the peripheral portion of that zone (Fig. 7b). In Figure 7f, where a much wider field was optically imaged than the standard field in order to cover the full extent of the activation, it is showed that the predominantly ventral portions of zones activated by stimulation at different frequencies overlapped extensively, as described previously. Interestingly, in accordance with this optically demonstrated overlap, anatomical overlap in projection from the two injection sites (the red circles in Fig. 7f) was found in the corresponding portion (dots in black pointed to by the arrowhead).

Discussion

Using OIS and binaural stimulation with trains of PT pips, we demonstrate a dorsoventral elongation of zone of activation, reminiscent of the isofrequency band (IFB). The zones of activation are more elongated along the dorsoventral axis for higher stimulus frequencies compared to the zones of activation for lower frequencies. The zones activated at different frequencies overlap extensively in their ventral portions, but less so centrally. The zone of activation corresponds well with electrophysiological responses, confirming the identification as IFB-like architecture. Multiunit recording reveals a relatively good, but not perfect, correspondence between the zones of activation and the distribution of matched electrophysiological responses. We also observe that the overall orientation of intrinsic axon terminals is aligned to the orientation of the zone of activation.

Comparison to Previous Imaging Studies

The global pattern of optically defined activation is highly reproducible. It should be noted, however, that the precise extent and contour of activated zones varied somewhat from session to session (compare the second with the third column in Fig. 2, and also see Fig. 5). This nonsystematic variation has also been noted in the rat barrel cortex (Masino and Frostig, 1996). We think that such fluctuations are not due to changes in the depth of anesthesia or in other physiological parameters. We found some variability in response at the threshold stimulation level. As shown in Figure 4, stimulation only at 8 kHz evoked virtually no optical response at the threshold SPL (40 dB) or only weak responses at a slightly higher SPL (50 dB). It is very unlikely that changes in the anesthetic level or heart rate would occur during the presentation of just one particular frequency. Rather, the failure to evoke optical signals fully is likely to reflect the fluctuation in response at threshold, and in this interpretation would faithfully represent the activation level under these imaging conditions. Masino and Frostig (1996) suggested that the variation in their responses might reflect a dynamically changing cortical function. It remains unclear, however, whether this is also the case for the experiments reported here. Since the wavelength of the illumination light is different (630 nm for them, and 540 nm in this study), the detected signal sources cannot be taken to be the same (Grinvald et al., 1999; Nemoto et al., 2004).

Previous OIS studies have investigated the tonotopic arrangement across the cat AI following either cochlear electrical stimulation (Dinse et al., 1997) or acoustic stimulation (Spitzer et al., 2001). In an experiment using acoustic stimulation, changes in light reflectance (darkening or activation) appeared to be widely spread across the AI; and areas of activation took, in general, an oval, rather than elongated, shape. Correlation between the optically and electrophysiologically detected responses was reported to be relatively weak, probably because of the extensive overlaps of the areas activated, even by PTs different by two octaves.

The methodologies employed in the present study and a similar study (Spitzer et al., 2001) are different in several respects, and this is likely to account for the different results obtained. Differences include the anesthetic agents and procedures, sound delivery styles, acoustic stimulation, and data analysis. First, we used isoflurane as an anesthetic agent. With this agent, the depth of anesthesia can be controlled so that it is maintained at a very stable level. Optical responses detected at a 540 nm wavelength are linked to blood volume changes (Grinvald et al., 1999; Sheth et al., 2003). Since anesthesia is known to strongly affect blood circulation (Eger, 1984), an unstable anesthetic level could cause fluctuations in the intensity and extent of optically detected activation. Such fluctuation would result in a mapping of less restricted activation areas, especially when utilizing an averaging process of many images. We went to great lengths to keep the heart rate and expiration CO2 level as stable as possible by frequently adjusting the anesthetic level during the experiment. Second, we delivered sound stimuli binaurally through a dichotic closed system, while Spitzer et al. (2001) adopted an open delivery system, with a speaker positioned at 50° in azimuth from the animal's midline and 25 cm distant from the ear contralateral to the imaged hemisphere. Because of the reflection and occlusion of the tones from the animal's pinnae and objects near the animal, this arrangement inevitably includes additional factors that could influence the temporal disparities and the level of the tones reaching the left and right ear (Reale and Kettner, 1986; Reale and Brugge, 1990; Zhang et al., 2004). This might well result in a different topography of activated neuronal clusters (Nakamoto et al., 2004). Third, the acoustic stimulus we used is a train of 10 repeats of PTs at a frequency of 5 Hz (Versnel et al., 2002), as opposed to the train of 5–7 PT repeats at a frequency of 1.25 Hz used by Spitzer et al. (2001). Their stimulus train was not only shorter but also less frequently presented. This variation is important because the responses of auditory cortical neurons in anesthetized animals are typically phasic in response to tone pips. In our own experience, where single bursts of PT pips were used as acoustic stimuli, just as they are used for electrophysiological determination of the response frequencies, they do not reliably evoke optical signals in the cat AI (data not shown). Furthermore, the optimal modulation rate for driving cat AI neurons with a train of PTs is known to be in the range of 5–8 Hz (Eggermont, 1991). This means that the number of repeats is critical to the reliable induction of blood volume changes at 540 nm (Nemoto et al., 2004). This is consistent with a previous study (Versnel et al., 2002) which has reported the optical imaging of tonotopy in ferrets. The repetition rate of ∼5–8 Hz is also comparable to the rate at which supragranular AI pyramidal neurons can generate action potentials without being affected by inhibitory postsynaptic potentials (at 50–150 ms duration) that follow the preceding excitatory postsynaptic potentials induced by PT stimulation (Ojima and Murakami, 2002).

Comparison to Previous Mapping Studies

The zones of activation induced by our stimulus set were dorsoventrally elongated or slightly tilted from this orientation (dorsoposterior–ventroanteriorly). This orientation is closely comparable to that of the IFBs defined by electrophysiological recording (Merzenich et al., 1975; Reale and Imig, 1980). The size is also comparable. The elongated IFB-like zones of activation are, on average, 5.6 mm long along the major axis for the higher frequencies (14 and 21 kHz), which is ∼1.4 times longer than that for the lower frequencies (4 and 8 kHz). This length is close to the averaged length (∼6 mm; see Reale and Imig, 1980) for the IF lines as defined electrophysiologically. It also corresponds to the dorsoventral extent of the activation regions demonstrated in an earlier optical recording in the cat (∼5.0 mm; Spitzer et al., 2001).

In our OIS results, lower frequencies of PT tended to produce less elongated zones of activation (3.6 mm long at 4 kHz and 4.4 mm long at 8 kHz; see Fig. 2). In the cat AI, it is known that lower frequencies (typically <4 kHz) are frequently mapped only in the anterior bank of the posterior ectosylvian sulcus (PES) along the dimension perpendicular to the cortical surface, and that IFBs for lower frequencies are localized only around the dorsal tip of the PES (Reale and Imig, 1980). This restricted representation of lower IFBs strongly corresponds to the less elongated activation detected by OIS around the dorsal tip of the PES following lower frequency stimulation.

Zones of activation induced by the stimulus set of PT pips showed a systematic shift across the cortical surface as a function of stimulus frequency. A close correspondence between the optical maps and multiunit recordings was found at 64% of the recording points that were located inside the boundary of at least one of the optically defined zones of activation (an ∼36% mismatch rate). Several factors may contribute to the less than perfect match. First, changes in light reflectance detected at 540 nm are known to reflect blood volume changes preferentially, and blood volume changes are linked only indirectly to neuronal activities (Grinvald et al., 1999; Sheth et al., 2003; Nemoto et al., 2004). Second, the partial correspondence between the two methods may result from scatter in the frequency response range at different cortical depths. For example, intermittent frequency gaps along a cortical column, where neurons cannot be effectively driven (Abeles and Goldstein, 1970), have been reported. Third, is the possible involvement of subthreshold neuronal activity in the induction of intrinsic signals (see below). This would lead to the failure of neurons to generate spikes in response to a stimulus frequency that evokes optical signals. Indeed, neurons in the overlap regions of multiple zones of activation frequently failed to respond to one of the multiple frequencies used as the effective stimuli (see above).

Possible Involvement of Subthreshold Neuronal Activities in Optical Responses

Many imaging studies in the visual and somatosensory systems have demonstrated good correspondence between neuronal firing (suprathreshold responses) and optically detected responses. This correspondence has been confirmed in three different respects. The first investigated simply whether the stimulus selectivity of neurons, recorded from usually one or a small number of sites within an optically defined response area, matched the stimulus used for optical imaging (Grinvald et al., 1986; Gochin et al., 1992; Shmuel and Grinvald 1996; Crair et al., 1997; Ghose and Ts'o, 1997; Maldonado et al., 1997; Shoham et al., 1997; Shoham and Grinvald, 2001; Tsunoda et al., 2001; Schummers et al., 2002). The second examined the correlation between the firing rate of neurons and the strength of optical signals (Shmuel and Grinvald 1996; Peterson et al., 1998). The third examined whether the spatial distribution of neurons exhibiting a particular stimulus selectivity corresponded to the extent of the optically detected activation (Masino et al., 1993; Roe and Ts'o, 1995; Masino and Frostig, 1996; Polley et al., 1999; Chen et al., 2001; Masino, 2003). Others have proposed that, in addition to the good correspondence for the suprathreshold response of neurons, extensive domains outside directly activated areas also exhibited increased optical responses, that is, subthreshold responses (Das and Gilbert, 1995; Toth et al., 1996).

In the auditory system, most previous studies have thus far regarded the contribution of subthreshold activities as unlikely. This is in part because of the good match between electrophysiologically detected suprathreshold activities and optically detected responses to the same tonal stimuli (see Bakin et al., 1996 for guinea pig; Harel et al., 2000 for chinchilla), and because of the relatively good match in spatial distribution of the two responses (Harrison et al., 2002). In the present study, however, we found the mismatch responses at nearly 35% of the recording points located within the optically defined zones of activation. That is, optically detected responsive zones included loci from which the corresponding multiunit activities are not recorded. This discrepancy may well be due to the contribution of subthreshold neuronal activities to optically detected responses.

Zones of Activation and Anatomical Intrinsic Connections

Injection of anterograde tracers was made within separate zones of activation demonstrated by OIS, either at their center or periphery. In either case, a close correspondence of the anatomical projection and optically demonstrated activation was observed in terms of the overall orientation and extent along the dorsoventral axis (see Fig. 7). In this respect, our results are consistent with previous studies (Matsubara and Phillips, 1988; Ojima et al., 1991; Wallace et al., 1991; Read et al., 2001). However, the results presented here show that the intrinsic projections originating from loci within a single activation zone distributed extensively within that zone (especially when injected in the center; see Fig. 7d,g,h). As shown in our electrophysiological mapping, the majority of these loci responded to the frequency used for visualizing that zone optically, despite having a variety of FRRs. Taken together, it seems likely that the intrinsic horizontal projections in the cat AI might interconnect domains representing similar FRRs (Read et al., 2001) rather than interconnecting domains showing similar best frequencies.

Superimposition of activation zones revealed optically at different stimulus frequencies demonstrates their overlap in the ventral segment of the imaged field or further ventral to it as well as, to a certain extent, in the dorsal segment. It strongly suggests that the region of overlap is outside AI or at the border between AI and AII of the cat auditory cortex (Schreiner and Cynader, 1984; Ojima and Takayanagi, 2004). Therefore, it is possible that neurons in the overlap regions are involved in processing spectrally complex tones that are comprised of multiple frequency components. This view is supported by the anatomical evidence for the convergence of projections originating from loci of sharply tuned multiple-unit responses in AI (Fig. 7f; see also Ojima and Takayanagi, 2004). In conclusion, neurons in the regions of overlap are likely to receive convergent synaptic input, each tuned to a distinct frequency, and can be regarded as one of the possible sites for the spectral, and probably temporal, integration of frequency components in complex tones. The specifics by which such integration is carried out and implications for an understanding of AI processing will have to be the subjects of future study.

We believe that the frequency representation demonstrated by the OIS will provide a fundamental framework for defining other acoustic parameters (Schreiner et al., 2000; Read et al., 2002) that are considered to be represented along and across the frequency gradient.

The authors are grateful to Drs M. Tanifuji, for continuous support during this experiment, K.S. Rockland, for critical reading of an earlier version of manuscript, and G. Uchida and T. Mogami, for advice in data analysis; and Mss M. Bellinger, K. Shirasawa and H. Mashiko for technical assistance in histology.

References

Abeles M, Goldstein MH Jr (
1970
) Functional architecture in cat primary auditory cortex: columnar organization and organization according to depth.
J Neurophysiol
 
33
:
172
–187.
Aitkin LM, Irvine DR, Nelson JE, Merzenich MM, Clarey JC (
1986
) Frequency representation in the auditory midbrain and forebrain of a marsupial, the northern native cat (Dasyurus hallucatus).
Brain Behav Evol
 
1929
:
17
–28.
Bakin JS, Kwon MC, Masino SA, Weinberger NM, Frostig RD (
1996
) Suprathreshold auditory cortex activation visualized by intrinsic signal optical imaging.
Cereb Cortex
 
6
:
120
–130.
Caird D, Scheich H, Klinke R (
1991
) Functional organization of auditory cortical fields in the Mongolian gerbil (Meriones unguiculatus): binaural 2-deoxyglucose patterns.
J Comp Physiol A
 
168
:
13
–26.
Chen LM, Friedman RM, Ramsden BM, LaMotte RH, Roe AW (
2001
) Fine-scale organization of SI (area 3b) in the squirrel monkey revealed with intrinsic optical imaging.
J Neurophysiol
 
86
:
3011
–3029.
Clarey JC, Barone P, Imig TJ (
1994
) Functional organization of sound direction and sound pressure level in primary auditory cortex of the cat.
J Neurophysiol
 
72
:
2383
–2405.
Cohen LB (
1973
) Changes in neuron structure during action potential propagation and synaptic transmission.
Physiol Rev
 
53
:
373
–418.
Crair MC, Ruthazer ES, Gillespie DC, Stryker MP (
1997
) Relationship between the ocular dominance and orientation maps in visual cortex of monocularly deprived cats.
Neuron
 
19
:
307
–318.
Das A, Gilbert CD (
1995
) Long-range horizontal connections and their role in cortical reorganization revealed by optical recording of cat primary visual cortex.
Nature
 
375
:
780
–784.
Dinse HR, Godde B, Hilger T, Reuter G, Cords SM, Lenarz T, von Seelen W (
1997
) Optical imaging of cat auditory cortex cochleotopic selectivity evoked by acute electrical stimulation of a multi-channel cochlear implant.
Eur J Neurosci
 
9
:
113
–119.
Eger EI 2nd(
1984
) The pharmacology of isoflurane.
Br J Anaesth
 
56
(Suppl 1):
71S
–99S.
Eggermont JJ (
1991
) Maturational aspects of periodicity coding in cat primary auditory cortex.
Hear Res
 
57
:
45
–56.
Formisano E, Kim DS, Di Salle F, van de Moortele PF, Ugurbil K, Goebel R (
2003
) Mirror-symmetric tonotopic maps in human primary auditory cortex.
Neuron
 
40
:
859
–869.
Frostig RD, Lieke EE, Ts'o DY, Grinvald A (
1990
) Cortical functional architecture and local coupling between neuronal activity and the microcirculation revealed by in vivo high-resolution optical imaging of intrinsic signals.
Proc Natl Acad Sci USA
 
87
:
6082
–6086.
Ghose GM, Ts'o DY (
1997
) Form processing modules in primate area V4.
J Neurophysiol
 
77
:
2191
–2196.
Gochin PM, Bedenbaugh P, Gelfand JJ, Gross CG, Gerstein GL (
1992
) Intrinsic signal optical imaging in the forepaw area of rat somatosensory cortex.
Proc Natl Acad Sci USA
 
89
:
8381
–8383.
Grinvald A, Lieke E, Frostig RD, Gilbert CD, Wiesel TN (
1986
) Functional architecture of cortex revealed by optical imaging of intrinsic signals.
Nature
 
324
:
361
–364.
Grinvald A, Shoham D, Shmuel D, Glaser I, Vanzetta E, Shtoyerman E, Slovin H, Sterkin C, Wijnbergen C, Hildesheim R and Arieli A (
1999
) In-vivo optical imaging of cortical architecture and dynamics. In: Modern techniques in neuroscience research (Windhorst U, Johansson H, eds), pp. 893–969. Berlin: Springer-Verlag.
Harel N, Mori N, Sawada S, Mount RJ, Harrison RV (
2000
) Three distinct auditory areas of cortex (AI, AII, and AAF) defined by optical imaging of intrinsic signals.
Neuroimage
 
11
:
302
–312.
Harrison RV, Harel N, Kakigi A, Raveh E, Mount RJ (
1998
) Optical imaging of intrinsic signals in chinchilla auditory cortex.
Audiol Neurootol
 
3
:
214
–223.
Harrison RV, Harel N, Panesar J, Mount RJ (
2002
) Blood capillary distribution correlates with hemodynamic-based functional imaging in cerebral cortex.
Cereb Cortex
 
12
:
225
–233.
Heil P, Rajan R, Irvine DR (
1992
) Sensitivity of neurons in cat primary auditory cortex to tones and frequency-modulated stimuli. II. Organization of response properties along the ‘isofrequency’ dimension.
Hear Res
 
63
:
135
–156.
Heil P, Rajan R, Irvine DR (
1994
) Topographic representation of tone intensity along the isofrequency axis of cat primary auditory cortex.
Hear Res
 
76
:
188
–120.
Hellweg FC, Koch R, Vollrath M (
1977
) Representation of the cochlea in the neocortex of guinea pigs.
Exp Brain Res
 
29
:
467
–474.
Hess A, Scheich H (
1996
) Optical and FDG mapping of frequency-specific activity in auditory cortex.
Neuroreport
 
7
:
2643
–2647.
Howard MA 3rd, Volkov IO, Abbas PJ, Damasio H, Ollendieck MC, Granner MA (
1996
) A chronic microelectrode investigation of the tonotopic organization of human auditory cortex.
Brain Res
 
724
:
260
–264.
Ikeda H, Wright MJ (
1974
) Sensitivity of neurones in visual cortex (area 17) under different levels of anaesthesia.
Exp Brain Res
 
20
:
471
–484.
Imig TJ, Adrian HO (
1977
) Binaural columns in the primary field (A1) of cat auditory cortex.
Brain Res
 
138
:
241
–257.
Kelly JB, Judge PW, Phillips DP (
1986
) Representation of the cochlea in primary auditory cortex of the ferret (Mustela putorius).
Hear Res
 
24
:
111
–115.
Kosaki H, Hashikawa T, He J, Jones EG (
1997
) Tonotopic organization of auditory cortical fields delineated by parvalbumin immunoreactivity in macaque monkeys.
J Comp Neurol
 
386
:
304
–316.
Krishna BS, Semple MN (
2000
) Auditory temporal processing: responses to sinusoidally amplitude-modulated tones in the inferior colliculus.
J Neurophysiol
 
84
:
255
–273.
Lütkenhöner B, Steinsträter O (
1998
) High-precision neuromagnetic study of the functional organization of the human auditory cortex.
Audiol Neurootol
 
3
:
191
–213.
Maldonado PE, Godecke I, Gray CM, Bonhoeffer T (
1997
) Orientation selectivity in pinwheel centers in cat striate cortex.
Science
 
276
:
1551
–1555.
Masino SA (
2003
) Quantitative comparison between functional imaging and single-unit spiking in rat somatosensory cortex.
J Neurophysiol
 
89
:
1702
–1712.
Masino SA, Frostig RD (
1996
) Quantitative long-term imaging of the functional representation of a whisker in rat barrel cortex.
Proc Natl Acad Sci USA
 
93
:
4942
–4947.
Masino SA, Kwon MC, Dory Y, Frostig RD (
1993
) Characterization of functional organization within rat barrel cortex using intrinsic signal optical imaging through a thinned skull.
Proc Natl Acad Sci USA
 
90
:
9998
–10002.
Matsubara JA, Phillips DP (
1988
) Intracortical connections and their physiological correlates in the primary auditory cortex (AI) of the cat.
J Comp Neurol
 
268
:
38
–48.
McMullen NT, Glaser EM (
1982
) Tonotopic organization of rabbit auditory cortex.
Exp Neurol
 
75
:
208
–220.
Merzenich MM, Brugge JF (
1973
) Representation of the cochlear partition of the superior temporal plane of the macaque monkey.
Brain Res
 
50
:
275
–296.
Merzenich MM, Knight PL, Roth GL (
1973
) Cochleotopic organization of primary auditory cortex in the cat.
Brain Res
 
63
:
343
–346.
Merzenich MM, Knight PL, Roth GL (
1975
) Representation of cochlea within primary auditory cortex in the cat.
J Neurophysiol
 
38
:
231
–249.
Merzenich MM, Kaas JH, Roth GL (
1976
) Auditory cortex in the grey squirrel: tonotopic organization and architectonic fields.
J Comp Neurol
 
166
:
387
–401.
Middlebrooks JC, Dykes RW, Merzenich MM (
1980
) Binaural response-specific bands in primary auditory cortex (AI) of the cat: topographical organization orthogonal to isofrequency contours.
Brain Res
 
181
:
31
–48.
Morel A, Garraghty PE, Kaas JH (
1993
) Tonotopic organization, architectonic fields, and connections of auditory cortex in macaque monkeys.
J Comp Neurol
 
335
:
437
–459.
Nakamoto KT, Zhang J, Kitzes LM (
2004
) Response patterns along an isofrequency contour in cat primary auditory cortex (AI) to stimuli varying in average and interaural levels.
J Neurophysiol
 
91
:
118
–135.
Nelken I, Bizley JK, Nodal FR, Ahmed B, Schnupp JW, King AJ (
2004
) Large-scale organization of ferret auditory cortex revealed using continuous acquisition of intrinsic optical signals.
J Neurophysiol.
 
92
:
2574
–2588.
Nemoto M, Sheth S, Guiou M, Pouratian N, Chen JW, Toga AW (
2004
) Functional signal- and paradigm-dependent linear relationships between synaptic activity and hemodynamic responses in rat somatosensory cortex.
J Neurosci
 
24
:
3850
–3861.
Ojima H, Murakami K (
2002
) Intracellular characterization of suppressive responses in supragranular pyramidal neurons of cat primary auditory cortex in vivo.
Cereb Cortex
 
12
:
1079
–1091.
Ojima H, Takayanagi M (
2001
) Use of two anterograde axon tracers to label distinct cortical neuronal populations located in close proximity.
J Neurosci Methods
 
104
:
177
–182.
Ojima H, Takayanagi M (
2004
) Cortical convergence from different frequency domains in the cat primary auditory cortex.
Neuroscience
 
126
:
203
–212.
Ojima H, Honda CN, Jones EG (
1991
) Patterns of axon collateralization of identified supragranular pyramidal neurons in the cat auditory cortex.
Cereb Cortex
 
1
:
80
–94.
Pantev C, Hoke M, Lehnertz K, Lutkenhoner B, Anogianakis G, Wittkowski W (
1988
) Tonotopic organization of the human auditory cortex revealed by transient auditory evoked magnetic fields.
Electroencephalogr Clin Neurophysiol
 
69
:
160
–170.
Peterson BE, Goldreich D, Merzenich MM (
1998
) Optical imaging and electrophysiology of rat barrel cortex. I. Responses to small single-vibrissa deflections.
Cereb Cortex
 
8
:
173
–183.
Polley DB, Chen-Bee CH, Frostig RD (
1999
) Two directions of plasticity in the sensory-deprived adult cortex.
Neuron
 
24
:
623
–637.
Ratzlaff EH, Grinvald A (
1991
) A tandem-lens epifluorescence macroscope: hundred-fold brightness advantage for wide-field imaging.
J Neurosci Methods
 
36
:
127
–137.
Read HL, Winer JA, Schreiner CE (
2001
) Modular organization of intrinsic connections associated with spectral tuning in cat auditory cortex.
Proc Natl Acad Sci USA
 
98
:
8042
–8047.
Read HL, Winer JA, Schreiner CE (
2002
) Functional architecture of auditory cortex.
Curr Opin Neurobiol
 
12
:
433
–440.
Reale RA, Brugge JF (
1990
) Auditory cortical neurons are sensitive to static and continuously changing interaural phase cues.
J Neurophysiol
 
64
:
1247
–1260.
Reale RA, Imig TJ (
1980
) Tonotopic organization in auditory cortex of the cat.
J Comp Neurol
 
192
:
265
–291.
Reale RA, Kettner RE (
1986
) Topography of binaural organization in primary auditory cortex of the cat: effects of changing interaural intensity.
J Neurophysiol
 
56
:
663
–682.
Roe AW, Ts'o DY (
1995
) Visual topography in primate V2: multiple representation across functional stripes.
J Neurosci
 
15
:
3689
–3715.
Sally SL, Kelly JB (
1988
) Organization of auditory cortex in the albino rat: sound frequency.
J Neurophysiol
 
59
:
1627
–1638.
Schreiner CE, Cynader MS (
1984
) Basic functional organization of second auditory cortical field (AII) of the cat.
J Neurophysiol
 
51
:
1284
–1305.
Schreiner CE, Sutter ML (
1992
) Topography of excitatory bandwidth in cat primary auditory cortex: single-neuron versus multiple-neuron recordings.
J Neurophysiol
 
68
:
1487
–1502.
Schreiner CE, Mendelson JR, Sutter ML (
1992
) Functional topography of cat primary auditory cortex: representation of tone intensity.
Exp Brain Res
 
92
:
105
–122.
Schreiner CE, Read HL, Sutter ML, Kilgard MP, Merzenich MM (
2000
) Modular organization of frequency integration in primary auditory cortex.
Annu Rev Neurosci
 
23
:
501
–529.
Schulze H, Hess A, Ohl FW, Scheich H (
2002
) Superposition of horseshoe-like periodicity and linear tonotopic maps in auditory cortex of the Mongolian gerbil.
Eur J Neurosci
 
15
:
1077
–1184.
Schummers J, Marino J, Sur M (
2002
) Synaptic integration by V1 neurons depends on location within the orientation map.
Neuron
 
36
:
969
–978.
Schwender D, Daunderer M, Klasing S, Finsterer U, Peter K (
1998
) Power spectral analysis of the electroencephalogram during increasing end-expiratory concentrations of isoflurane, desflurane and sevoflurane.
Anaesthesia
 
53
:
335
–342.
Sheth S, Nemoto M, Guiou M, Walker M, Pouratian N, Toga AW (
2003
) Evaluation of coupling between optical intrinsic signals and neuronal activity in rat somatosensory cortex.
Neuroimage
 
19
:
884
–894.
Sheth S, Nemoto M, Guiou M, Walker M, Pouratian N, Toga AW (
2004
) Linear and nonlinear relationships between neuronal activity, oxygen metabolism, and hemodynamic responses.
Neuron
 
42
:
347
–355.
Shmuel A, Grinvald A (
1996
) Functional organization for direction of motion and its relationship to orientation maps in cat area 18.
J Neurosci
 
16
:
6945
–6964.
Shoham D, Grinvald A (
2001
) The cortical representation of the hand in macaque and human area S-I: high resolution optical imaging
J Neurosci
 
21
:
6820
–6835.
Shoham D, Hubener M, Schulze S, Grinvald A, Bonhoeffer T (
1997
) Spatio-temporal frequency domains and their relation to cytochrome oxidase staining in cat visual cortex.
Nature
 
385
:
529
–533.
Spitzer MW, Calford MB, Clarey JC, Pettigrew JD, Roe AW (
2001
) Spontaneous and stimulus-evoked intrinsic optical signals in primary auditory cortex of the cat.
J Neurophysiol
 
85
:
1283
–1298.
Talavage TM, Sereno MI, Melcher JR, Ledden PJ, Rosen BR, Dale AM (
2004
) Tonotopic organization in human auditory cortex revealed by progressions of frequency sensitivity.
J Neurophysiol
 
91
:
1282
–1296.
Taniguchi I, Horikawa J, Moriyama T, Nasu M (
1992
) Spatio-temporal pattern of frequency representation in the auditory cortex of guinea pigs.
Neurosci Lett
 
146
:
37
–40.
Thomas H, Tillein J, Heil P, Scheich H (
1993
) Functional organization of auditory cortex in the mongolian gerbil (Meriones unguiculatus). I. Electrophysiological mapping of frequency representation and distinction of fields.
Eur J Neurosci
 
5
:
882
–897.
Toth LJ, Rao SC, Kim DS, Somers D, Sur M (
1996
) Subthreshold facilitation and suppression in primary visual cortex revealed by intrinsic signal imaging.
Proc Natl Acad Sci USA
 
93
:
9869
–9874.
Tsunoda K, Yamane Y, Nishizaki M, Tanifuji M (
2001
) Complex objects are represented in macaque inferotemporal cortex by the combination of feature columns.
Nat Neurosci
 
4
:
832
–838.
Tsytsarev V, Tanaka S (
2002
) Intrinsic optical signals from rat primary auditory cortex in response to sound stimuli presented to contralateral, ipsilateral and bilateral ears.
Neuroreport
 
13
:
1661
–1666.
Uno H, Murai N, Fukunishi K (
1993
) The tonotopic representation in the auditory cortex of the guinea pig with optical recording.
Neurosci Lett
 
150
:
179
–182.
Versnel H, Mossop JE, Mrsic-Flogel TD, Ahmed B, Moore DR (
2002
) Optical imaging of intrinsic signals in ferret auditory cortex: responses to narrowband sound stimuli.
J Neurophysiol
 
88
:
1545
–1558.
Villeneuve MY, Casanova C (
2003
) On the use of isoflurane versus halothane in the study of visual response properties of single cells in the primary visual cortex.
J Neurosci Methods
 
129
:
19
–31.
Wallace MN, Kitzes LM, Jones EG (
1991
) Intrinsic inter- and intralaminar connections and their relationship to the tonotopic map in cat primary auditory cortex.
Exp Brain Res
 
86
:
527
–544.
Winer JA (
1984
) The pyramidal neurons in layer III of cat primary auditory cortex (AI).
J Comp Neurol
 
229
:
476
–496.
Woolsey CN, Walzl EM (
1942
) Topical projection of nerve fibers from local regions of the cochlea to the cerebral cortex of the cat.
Bull John Hopoins Hosp
 
71
:
315
–344.
Zhang J, Nakamoto KT, Kitzes LM (
2004
) Binaural interaction revisited in the cat primary auditory cortex.
J Neurophysiol
 
91
:
101
–117.

Author notes

1Cortical Organization Systematics, BSI, RIKEN, 2-1 Hirosawa, Wako, Saitama, 351-0198 Japan, 21st Department of Anatomy, Toho University School of Medicine, Toho, 5-21-16, Omori Nishi, Ota-ku, Tokyo, 143-8540, Japan and 3Integrative Neural Systems, BSI, RIKEN, 2-1 Hirosawa, Wako, Saitama, 351-0198 Japan