Research report N euronal activity in the inferior

Brain Research 979 (2003) 153–164 www.elsevier.com / locate / brainres

Research report

Neuronal activity in the inferior colliculus and bordering structures during vocalization in the squirrel monkey * ¨ Florian Pieper, Uwe Jurgens ¨ , Germany German Primate Center, Kellnerweg 4, 37077 Gottingen Accepted 24 April 2003

Abstract In four squirrel monkeys (Saimiri sciureus), the inferior colliculus, together with the neighboring superior colliculus, reticular formation, cuneiform nucleus and parabrachial area, were explored with microelectrodes, looking for neurons that might be involved in the discrimination between self-produced and external sounds. Vocalization was elicited by kainic acid injections into the periaqueductal gray of the midbrain. Acoustic tests were carried out with ascending and descending narrow-band noise sweeps spanning virtually the whole hearing range of the squirrel monkey. Altogether 577 neurons were analyzed. Neurons that both were audiosensitive and fired in advance of self-produced vocalization were found almost exclusively in the pericentral nuclei of the inferior colliculus and the adjacent reticular formation. Only the latter, however, contained, in addition, neurons that fired during external acoustic stimulation, but remained quiet during self-produced vocalization. These findings suggest that the reticular formation bordering the inferior colliculus is involved in the discrimination between self-produced and foreign vocalization on the basis of a vocalmotor feedforward mechanism.  2003 Elsevier B.V. All rights reserved. Theme: Neural basis of behavior Topic: Neuroethology Keywords: Inferior colliculus; Reticular formation; Vocalization; Squirrel monkey; Extracellular recording

1. Introduction Squirrel monkeys, like numerous other species, often respond to vocalizations of conspecifics with vocalizations of the same type with very short latency (100–200 ms) [14]. As the animals never echo their own self-produced vocalizations, it must be assumed that there is a mecha-

nism which allows the animal to distinguish self-produced from external vocalizations within very short time. One possible way to master this problem is to send a corollary discharge of the vocal motor command to the auditory system, informing the latter about the forthcoming acoustic event. The present study is an attempt to find out whether the inferior colliculus and / or bordering tegmentum contain

Abbreviations: Aq, aquaeductus cerebri; BC, brachium conjunctivum; BCI, brachium colliculi inferioris; BP, brachium pontis; CCI, commissura colliculi inferioris; CCS, commissura colliculi superioris; IC, colliculus inferior; ICc, colliculus inferior, nucl. centralis; ICd, colliculus inferior, nucl. dorsalis; ICx, colliculus inferior, nucl. externus; CP, commissura posterior; SC, colliculus superior; SCi, colliculus superior, stratum intermediale; SCp, colliculus superior, stratum profundum; SCs, colliculus superior, stratum superficiale; CT, corpus trapezoideum; Cun, nucl. cuneiformis; DCN, nucl. cochlearis dorsalis; DG, nucl. dorsalis tegmenti; FLM, fasciculus longitudinalis medialis; FRM, formatio reticularis myelencephali; FRPc, formatio reticularis pontis caudalis; FRPo, formatio reticularis pontis oralis; FRTM, formatio reticularis tegmenti mesencephali; GPo, griseum pontis; IV, 4th ventricle; LC, locus coeruleus; LCb, lingula cerebelli; LL, lemniscus lateralis; LLd, nucl. lemnisci lateralis dorsalis; LLv, nucl. lemnisci lateralis ventralis; LM, lemniscus medialis; Mv, nucl. motorius n. trigemini; Nab, nucleus ambiguus; NCT, nucl. trapezoidalis; Niv, nucl. trochlearis; NMv, nucl. tractus mesencephalicus n. trigemini; NSv, nucl. tractus spinalis n. trigemini; nvi, nervus abducens; Nvi, nucl. abducens; Nvii, nucl. facialis; nvm, nervus trigeminus major; nvn, nervus trigeminus minor; OI, oliva inferior; OS, oliva superior; PAG, periaqueductal gray; PAGd, dorsal periaqueductal gray; PAGv, ventral periaqueductal gray; PbL, nucl. parabrachialis lateralis; PBM, nucl. parabrachialis medialis; PL, paralemniscal zone; Pg, nucl. parabigeminalis; Pv, nucl. principalis n. trigemini; Py, tractus pyramidalis; RTP, nucl. reticularis tegmenti pontis; Sag, sagulum; TMv, tractus mesencephalicus n. trigemini; TTp, tractus tectopontinus; TTs, tractus tectospinalis; VCN, nucl. cochlearis ventralis; VMA, velum medullare anterius; VT, nucl. ventralis tegmenti *Corresponding author. Tel.: 149-551-385-1250; fax: 149-551-385-1302. ¨ E-mail address: [email protected] (U. Jurgens). 0006-8993 / 03 / $ – see front matter  2003 Elsevier B.V. All rights reserved. doi:10.1016 / S0006-8993(03)02897-X

154

¨ / Brain Research 979 (2003) 153–164 F. Pieper, U. Jurgens

neurons which, on the one hand, receive an auditory and vocalmotor input and, on the other hand, react differently to self-produced and external sounds. The inferior colliculus was chosen because of the following reasons: (1) the inferior colliculus is an obligatory relay station of the central auditory pathway in the majority of mammals [1]; (2) the inferior colliculus contains three subnuclei, two of which (nucl. externus and nucl. dorsalis), because of their multisensory and corticofugal input, are predisposed to act as higher-order auditory control centers [8,12,13]; (3) The nucl. externus of the inferior colliculus receives a direct input from the periaqueductal gray, the most important vocalization-control structure in the midbrain [6,18]. The tegmentum bordering the inferior colliculus was chosen because it also receives auditory as well as periaqueductal inputs [23,29]. The working hypothesis underlying the present study may be formulated in the following way: If there is a brain region discriminating external and selfproduced vocalization on the basis of a vocalmotor feedforward mechanism, such a region should contain neurons that, at the same time, are sensitive to external acoustic stimuli and change their activity in advance of self-produced calls. The present study is a search for such neurons in the inferior colliculus and bordering tegmentum, using the extracellular single-unit recording technique.

2. Material and methods The experiments were carried out in four adult squirrel monkeys (Saimiri sciureus, one male, three females). After having familiarized the animals with the daily handling procedures and sitting in the restraining chair, a recording chamber was implanted onto the skull in a stereotaxic operation under deep anesthesia (35 mg / kg pentobarbital sodium). The recording chamber consisted of a plexiglas cylinder (16 mm i.d.) with a platform (40340 mm) for head fixation. Within the cylinder, there was a narrow bridge of dental acrylic (Paladur) near the midsagittal plane, containing 10 guiding tubes (0.5 mm i.d.). Electrical stimulation electrodes as well as pharmacological injection cannulae were introduced into the brain via these guiding tubes. The platform was fixed on the skull by the aid of five stainless steel screws, anchored in the bone with nuts and dental cement. Three screws were interconnected by a thin silver-wire with a plug-in connector and used as indifferent electrode. The bone underneath the lumen of the cylinder was removed, leaving the dura mater intact. The dura and inner surface of the cylinder were cleaned at the beginning of each recording session. In periods without recording, cleaning took place at least every second day. Between the experiments, the cylinder was covered with a lid. The platform could be fixed to a holder mounted above the monkey chair to prevent animal head movements during recording.

Ten days after platform fixation, an electrical stimulation electrode was implanted into the periaqueductal gray of the midbrain (PAG). The electrode consisted of a stainless steel tube of 0.47 mm outer diameter and a central teflon-coated stainless steel wire that protruded 2 mm from the tube. The wire was uninsulated at its tip for 1 mm. The electrode was lowered into the brain stem in steps of 0.5 mm. After each step, the electrical elicitability of vocalization was tested, using trains of biphasic rectangular pulses (1 ms pulse duration, 33 Hz pulse rate, 40–300 mA peak current and 1–10 s train duration). If a site within the PAG was found at which the animal vocalized in a relaxed manner with a stimulus current below 100 mA, the electrode was fixed to the platform with dental cement. For pharmacological stimulation, the core wire of the electrode was removed and replaced temporarily by a cannula (gauge 33) connected to a Hamilton microsyringe with a flexible teflon tube. To test whether or not a site producing vocalization with electrical stimulation also produces vocalization with pharmacological stimulation, 0.05–0.2 ml homocysteic acid (2–10 mg / ml) was injected. If homocysteic acid produced vocalization, this site was used in the following recording sessions for the elicitation of vocalization. In the recording sessions, however, kainic acid (0.1–0.2 ml, 25–250 ng / ml) was used instead of homocysteic acid, as kainic acid-induced vocal sequences lasted much longer (ca. 20 min) than homocysteic acidinduced sequences (ca. 2 min). As kainic acid has a neurotoxic effect, the elicitability of vocalization decreased over time. To compensate for the shortening of the vocal sequences, the dose was increased from time to time, so that per session two sequences of 15–20 min each were obtained. If no long-lasting vocalization sequences could be obtained anymore, a new stimulation electrode was implanted at another position within the PAG. The neuronal activity was recorded with epoxy-insulated tungsten microelectrodes (250 mm shaft diameter) with an impedance of 5 MV (A-M-Systems; [5755). The microelectrodes were introduced stereotaxically into the brain using a calibrated xyz-micropositioner of 25 mm travel range in each direction. On this micropositioner, a computer controlled Kopf microdrive (Type 640) was mounted exactly in parallel to the positioner’s z-axis. The electrode position was determined before and after each penetration relative to a reference point on the recording chamber. The neuronal signal was fed into a high impedance pre-amplifier stage (NL 100, Digitimer), filtered (300 Hz–3 kHz) and amplified by standard equipment to a 1 V (peak-topeak) level. Vocalizations were recorded with a Sennheiser directional microphone (MKH 416) placed about 80 cm in front of the animal’s mouth. The signal was filtered and amplified to a 1 V level. The microphone signal as well as the neuronal signal were digitized with a NI-A2100 analog I / O board (sampling rate: 48 kHz / 16 bit) in a Macintosh Computer. The custom-written ‘C’-program for harddisk-


recording was also used to control the Kopf microdrive and save a protocol of the recording positions. The recording electrode was positioned at a site just above the inferior colliculus. After starting the pharmacological stimulation, the electrode was lowered in steps of 40 mm. If a site was found showing good signal / noise ratio, recordings were taken during the production of 20–50 vocalizations. The time between single vocalizations was used to test the neuronal reaction to external acoustic stimuli. Linear narrow-band noise sweeps (center frequency: 200 Hz to 20 kHz, resp. 20 kHz to 200 Hz; bandwidth 316 Hz; rise and fall times 30 ms) were ¨ and Kjaer sine / noise generator generated with a Bruel (Type 1049). In these tests, 12 rising and 12 falling sweeps alternated with each other, each sweep lasting 1 s and being separated from the next by 500 ms. Only those sweeps were evaluated which had a minimal time interval of 200 ms to the preceding vocalization and did not overlap with the following vocalization. For determination of the exact starting time of the acoustical stimulation, a TTL signal was fed into the microphone channel with each sweep. The sweeps were amplified with a Sony F335R HiFi-amplifier and played to the animal via a ‘Teufel 200’ loudspeaker positioned 2 m in front of the animal’s head in a sound-attenuated chamber. The loudspeaker had a frequency response of 61 dB from 60 Hz to 20 kHz, a signal-to-noise ratio of .90 dB and a distortion of ,0.1%. The sound-pressure level was 5761 dB, measured at the animal’s ear with a Rhode and Schwarz sound level meter (EGT 201.4318). This relatively low level was used to avoid restlessness of the animal which regularly appeared during the presentation of higher intensity sounds. After all tests had been completed at a particular site, recording was stopped and a new site was searched that was at least 160 mm below the former. This procedure was repeated until the animal stopped to vocalize. Recording tracks were spaced 0.4 mm apart in the medio-lateral direction and 1 mm in the antero-posterior direction. Small electrolytic lesions (10 mA, 20 s, tip negative) were placed by the aid of a custom-made direct current source at the end of some of the electrode tracks for better identification of the electrode positions. Up to 18 electrode tracks were explored per animal. Only one track was explored per day, in order to keep the session time at or below 2 h. At the end of the experiment, the animals were killed with an overdose of pentobarbital sodium and perfused through the heart with heparinized warm physiological saline followed by 4% paraformaldehyde in phosphatebuffered saline (pH 7.4). The brain was removed from the skull in blocks of 6–10 mm thickness, with surfaces cut in the stereotaxic AP-plane. For 1 week, the blocks were stored in 4% paraformaldehyde. A few days before cutting, the blocks were transferred to 20% sucrose. Sections were cut at 40 mm and mounted on slides. Alternating sections underwent Nissl-staining with cresyl violet and immunohistological glial fibrillary acid protein-staining [5] for

155

track identification, respectively. Enlarged drawings (1:10) of the sections were made for reconstruction of the electrode tracks and marking lesions. These drawings, together with the position protocols of the electrodes, were used to determine the recording sites. A custom-written PPC LabView application was used for isolation of single units with a window discriminator function as well as data analysis. First, for all elicited vocalizations and acoustic sweeps, respectively, exact onset times as well as call type and sweep type were determined. Then, peri-event time histograms of the spike responses were calculated for all vocalizations and sweeps of the same type recorded at a specific site. The experiments were approved by the animal ethics committee of the district government of Braunschweig, Lower Saxony.

3. Results All vocalization-eliciting stimulation sites were located in the PAG between 0 and 1.2 mm lateral to the sagittal midline and between 0 and 2.0 mm anterior to the stereotaxic AP0 plane, according to the stereotaxic atlas of Emmers and Akert [11]. Virtually all stimulation sites yielded mixed vocal sequences, consisting of more than one call type. About half of the elicited calls were ‘cawing’ calls. Cawing calls are harmonically structured calls with a fundamental frequency of 200–700 Hz and numerous overtones; the frequency range spans about 6 kHz. About 40% of the elicited calls represented ‘cackling’, ‘trilling’ and ‘chuck’ calls. These calls are characterized by a steep frequency modulation of the fundamental over several kHz. In the case of the cackling and trilling calls, this frequency modulation is rhythmic, with a repetition rate of ca. 13 Hz. The frequency range (including harmonics) is about 20 kHz in the case of cackling and 8 kHz in the case of trilling. The remaining ca. 10% of elicited calls were made up of ‘peeping’ calls (fundamental frequency: 5–10 kHz), ‘shrieking’ calls (non-harmonic, i.e. noise-like calls with a frequency range of more than 10 kHz) and ‘yapping’ calls (short explosive calls, consisting of a steep, frequency-modulated component and a wide-frequency noise-like component). The recording positions are shown in Fig. 1a–f (middle upper diagrams). They are distributed throughout the dorsolateral midbrain / pons transitional area, including the inferior and superior colliculi, dorsolateral mesencephalic and paralemniscal pontine reticular formation, lateral periaqueductal gray, cuneiform nucleus, parabrachial region, locus coeruleus and mesencephalic trigeminal nucleus. Altogether 577 positions were analyzed. Only those positions entered analysis which could be tested with at least six vocalizations of the same type. Not all of these positions, however, could be tested acoustically. In 46% of the positions, the pharmacologically elicited vocalizations

156


Fig. 1. (a) Frontal sections of the squirrel monkey’s brain with distribution of recording sites. Upper left: stereotaxic plane with labeling of brain structures (see Abbreviations). Upper middle: Positions of all recording sites tested at that plane. Upper right: Position of cells starting to fire before vocalization onset. Squares indicate audio-sensitive cells; dots indicate cells unresponsive to acoustic stimuli or being untested acoustically. Lower left: Position of cells increasing their activity after vocalization onset. Lower middle: Position of cells increasing their activity to external acoustic stimuli, but not to self-produced vocalization. Lower right: Position of cells increasing their activity after vocalization onset, but being unresponsive to the acoustic test stimuli. (b–f) For explanation, see (a).

followed each other so fast, that it was not possible to run a complete acoustic test (12 upward and 12 downward sweeps) in the periods between single calls of the vocal sequence. In Table 1, therefore, the results are presented separately for those positions tested acoustically and those tested with self-produced vocalizations only. Table 1 shows that out of the 577 analyzed recording positions, 35 exhibited an increase in activity in advance of self-produced vocalizations. A major part of these cells (16) lay in the lateral pontine reticular formation, some of

them close to the ventral nucleus of the lateral lemniscus, others following the course of the lateral lemniscus dorsally, with some located even within the lateral lemniscus (Fig. 1c). Eight of them could be tested acoustically. None responded to acoustic stimuli. In the inferior colliculus and immediately bordering mesencephalic reticular formation, 13 neurons were found to fire before vocalization onset. Ten of them were tested acoustically. Of these, nine responded to external acoustic stimuli. All inferior colliculus neurons showing pre-onset activity were located in


157

Fig. 1b

the dorsal and external nuclei; none was found in the central nucleus. Neurons firing before vocalization onset were also found in the cuneiform nucleus (Fig. 2). None of these neurons reacted to acoustic stimuli. With the exception of one cell in the dorsal nucleus of the inferior colliculus and one cell in the caudolateral pontine reticular formation, all pre-onset neurons showed an increased activity not only just before vocalization, but also during vocalization. The latency between activity increase and vocalization onset of pre-onset neurons varied between 10 and 200 ms. No statistical differences were found between different structures. Out of the 577 neurons tested, 352 showed a voc-

alization-correlated activity starting after vocalization onset (Table 1). Neurons of this type were mainly found in the inferior colliculus and bordering mesencephalic reticular formation, lateral pontine reticular formation, cuneiform nucleus and between the fiber bundles of the lateral lemniscus. If the ratio is calculated between audio-sensitive cells and cells not responding to acoustic stimuli, characteristic differences appear between the different brain structures. In the central nucleus of the inferior colliculus, 93% of the neurons being activated during vocalization reacted also to external acoustic stimuli. In the pericentral nuclei, nucl. externus and dorsalis, only 63% of the cells reacted to the acoustic stimuli. In the lateral

158


Fig. 1c

reticular formation, a systematic rostrocaudal trend was found with 81% of the mesencephalic peri-collicular neurons reacting to acoustic stimuli, 45% of the oral pontine and 0% of the caudal pontine reticular cells. The cuneiform nucleus, which lies at the border between midbrain and pons, contained 41% audio-sensitive cells. A number of cells, being active during vocalization, changed their discharge rate with changes in the fundamental frequency (only cells that could be tested with cackling and trilling calls, that is, rhythmically frequencymodulated calls, were used for analysis in this case). Cells with frequency-dependent discharge rates were found in a greater number (more that 10) only in the central nucleus

of the inferior colliculus, the mesencephalic reticular formation bordering the inferior colliculus, the cuneiform nucleus, in the lateral lemniscus itself and in the lateral reticular formation of the caudal pons. While almost all cells in the central nucleus of the inferior colliculus, mesencephalic reticular formation and lateral lemniscus reacted to external acoustic stimuli in a frequency-specific manner, only half of the nucl. cuneiformis cells and none of the caudal pontine reticular cells did so (Table 1). Neurons that did not react to the acoustic test stimuli, but increased their activity during self-produced vocalization, were found mainly in the caudolateral pontine reticular formation (18), cuneiform nucleus (10), dorsola-


159

Fig. 1d

teral mesencephalic reticular formation (9) and dorsal nucleus of the inferior colliculus (9). One hundred and eighty-eight of the 577 recorded neurons did not increase their activity during vocalization, and 103 of them were tested acoustically. Of these, nine reacted to external acoustic stimuli. As the spontaneous activity in all nine neurons was very low (less than 1 spike per 3 s), it was not possible to decide whether or not there was a decrease in activity during vocalization and, if yes,

whether this decrease began before or after vocalization onset. The fact, however, that these cells increased their activity to external acoustic stimuli, but not to self-produced acoustic stimuli, suggests that these cells are either directly inhibited during vocalization or are detached from their normal acoustic input by inhibitory processes acting on their input neurons. Neurons that did not react to self-produced vocalization, but reacted to external acoustic stimuli, were found in the mesencephalic reticular forma-

160


Fig. 1e

tion bordering the inferior colliculus (4), in the central (3) and dorsal nucleus (1) of the inferior colliculus and in the stratum profundum of the superior colliculus (1).

4. Discussion The results of the present study make clear that the various structures of the lateral midbrain / pons region are involved in audio-vocal processes in very different ways.

There are structures which contain a relatively high percentage of cells which are activated during self-produced vocalization, but not during listening to external acoustic stimuli. The most prominent example is the caudolateral pontine reticular formation. The periaqueductal gray shows a similar tendency. Both structures also contain cells which increase their activity before vocalization onset. These observations suggest that the caudolateral pontine reticular formation and periaqueductal gray are concerned with vocal production rather than auditory


161

Fig. 1f

perception. The fact that the caudolateral pontine reticular formation contains a relatively high number of cells which change their discharge rate phase-correlated to the frequency modulation of the vocalization, whereas the periaqueductal gray seems to lack such neurons completely (see also Ref. [10]), furthermore, suggests that the caudolateral pontine reticular formation is involved in vocal pattern control, while the periaqueductal gray rather serves gating functions. This interpretation is supported by the

observation that electrical stimulation of the caudolateral pontine reticular formation produces abnormal vocalization, while stimulation of the periaqueductal gray yields natural calls [15,17]. The finding that lesions in the caudolateral pontine reticular formation affect call structure, while lesions in the periaqueductal gray abolish specific call types completely, point into the same direction [16,18,19]. A gating function of the periaqueductal gray is also suggested by the fact that it contains neurons which


162

Table 1 Brain area-specific vocalization-related neuronal activity Structure

Pre-onset activity Cells record.

ICc ICd ICx SCi SCp Cun FRTM FRPo FRPc FRM LL LM BC PAG NMv LC PbL PbM

0 3 2 0 1 3 8 0 13 0 3 0 0 1 0 1 0 0

Post-onset activity Audio test pos.

neg.

0 2 1 0 0 0 6 0 0 0 0 0 0 0 0 0 0 0

0 1 0 0 0 3 0 0 6 0 2 0 0 0 0 0 0 0

Cells record.

59 20 13 0 4 16 43 16 8 2 11 4 13 6 1 0 2 2

Audio test pos.

neg.

32 4 7 0 1 5 16 4 0 0 6 3 0 1 0 0 0 2

2 4 1 0 1 5 6 5 1 2 1 0 3 2 0 0 0 0

Frequency-modulated activity

No activity during vocalization

Cells record.

Cells record.

25 6 3 0 2 14 28 6 24 0 11 5 6 0 2 0 0 0

Audio test pos.

neg.

9 1 2 0 1 2 19 1 0 0 7 3 4 0 0 0 0 0

1 4 0 0 0 2 3 1 11 0 1 0 0 0 2 0 0 0

8 16 7 4 19 4 51 23 11 4 1 0 24 10 0 2 2 2

Audio test pos.

neg.

3 1 0 0 1 0 4 0 0 0 0 0 0 0 0 0 0 0

4 13 3 4 15 2 22 8 4 4 1 0 9 2 0 0 1 2

Pre-onset activity, cells starting to fire before vocalization onset; Post-onset activity, cells firing after vocalization onset; Frequency-modulated activity, cells changing their discharge rate in the rhythm of the frequency modulation during cackling and trilling calls; No activity during vocalization, cells that do not increase their discharge rate during vocalization; Audio test pos., cells increasing their activity during the acoustic test stimuli; Audio test neg., cells unresponsive to the acoustic test stimuli; Cells record., number of cells recorded (the difference between this number and the sum of ‘Audio test pos.’ and ‘Audio test neg.’ cells gives the number of recorded cells not tested with external acoustic stimuli). Structure: see Abbreviations.

exclusively react to vocalization-eliciting acoustic stimuli. These neurons do not react to acoustic stimuli that do not induce vocalization. When they react, they fire throughout the interval between acoustic stimulus and vocal reaction [10]. A behavior very different from that of the caudolateral pontine reticular and periaqueductal neurons is shown by neurons of the central nucleus of the inferior colliculus. In contrast to the first two structures, almost all cells of the central nucleus of the inferior colliculus that were active during vocalization also reacted to acoustic stimuli (93%). Not a single cell with vocalization-correlated activity started to fire before vocalization onset. Furthermore, none of the neurons was active during the complete duration of the tone sweep. Activity was always limited to a certain frequency range. The tonotopy corresponded to that described by FitzPatrick [12]. These findings are in harmony with the well-known role of the central nucleus as a classical relay station of the auditory pathway. Only 7% of the cells with vocalization-correlated activity did not react to the acoustic test stimuli. It is unclear whether the lack of reaction in these cases was due to an intensity too low to activate the neurons, or to complex response characteristics not matched by the test tones, or to the neurons’ general unresponsiveness to acoustic stimuli. Behavioral tests have shown that the auditory threshold of the squirrel monkey is much below the intensity level used in the present study: the threshold is 48 dB at 200 Hz and 18 dB at 20 kHz, with a minimum of 7 dB at 4 kHz [32]. The high

percentage of audio-positive cells found in the central nucleus indicates that the acoustic test stimuli used in the present study were effective in activating at least the vast majority of auditory-responsive cells. The pericentral nuclei of the inferior colliculus, that is, nucl. externus and dorsalis, differed in their responses in several respects from the central nucleus. While in the latter, almost all recorded neurons were activated by the auditory test stimuli, less than half of the cells in the pericentral nuclei were audio-sensitive. A possible explanation for this difference is provided by neuroanatomical studies. These studies show that in contrast to the central nucleus, the pericentral nuclei are not purely auditory, but receive, in addition, somatosensory input from the spinal cord, dorsal column nuclei, spinal trigeminal nucleus and primary somatosensory cortex [9,30,31]. Differences in post-stimulus inhibition cannot account for the response differences between central and pericentral nuclei, as poststimulus inhibitions usually do not last longer than 100 ms, while the interval between self-produced vocalization and acoustic stimuli was at least 200 ms [3,7,22]. Another difference between pericentral and central nuclei found in the present study relates to the relative number of frequency-modulated cells, that is, cells changing their discharge rate in relation to changes in fundamental frequency of vocalization. The central nucleus contains about 50% more frequency-modulated cells than the pericentral nuclei. This difference is probably due to two facts. One is that the tuning curves of the pericentral neurons, in


Fig. 2. Peri-event time histogram (a) and spike recording (b) from the cuneiform nucleus during vocalization. Arrows indicate onset of vocalization-correlated activity. The peri-event time histogram is based on the production of 12 ‘chuck’ calls, starting at time 0.0; bin width: 5 ms. The traces in (b) represent the synchronized neuronal activity, microphone signal and sonagram during the production of a single ‘chuck’.

general, are broader and more irregular than those found in the central nucleus [2]. Secondly, the pericentral neurons typically show a rapid adaptation to repetitive stimuli [1]. As in the present study, frequency sensitivity was determined on whether or not neurons changed their discharge rate in the rhythm of rhythmically frequency-modulated vocalization, the repetitive character of the frequency-modulation might have concealed any frequencydependence. The most prominent difference found between pericentral and central nuclei relates to the number of pre-onset neurons. While we could not find a single neuron in the central nucleus firing in advance of vocalization onset, five of such neurons were found in the pericentral nuclei. Four of them could be tested acoustically. Three of these neurons were audio-sensitive, and thus, according to the criteria outlined in the Introduction, might contribute to the distinction between self-produced and externally produced vocalizations on the basis of feedforward mechanisms. The external and dorsal nuclei of the inferior colliculus are not the only areas containing audio-sensitive, vocalization-correlated neurons with pre-onset activity.

163

Another one is the mesencephalic reticular formation bordering the inferior colliculus rostroventrally. Out of the 79 neurons with vocalization-correlated activity in this area, eight showed a pre-onset activity. Six of these neurons could be tested acoustically. All of them were audio-sensitive. The main difference between the pericentral nuclei of the inferior colliculus and the reticular formation was that the latter contained a higher number of neurons that increased their activity to external acoustic stimuli, but not to self-produced vocalization. In the external and dorsal nuclei of the inferior colliculus together, only one out of 17 neurons (5.9%) showed this behavior, while in the reticular formation four out of 26 neurons (15%) behaved in this way. The mesencephalic reticular formation thus is the only structure found in this study in which auditory input and vocalmotor output information converges, and where a greater number of cells is found reacting differently to self-produced and externally produced auditory input. The reticular formation bordering the inferior colliculus thus, even more than the inferior colliculus itself, fulfills the criteria for a vocalmotor feedforward mechanism capable of distinguishing self-produced and external vocalizations. The anatomical basis for a convergence of vocalmotor and auditory information is given on the one hand by a direct projection of the periaqueductal gray into the reticular formation bordering the inferior colliculus [23]. The periaqueductal gray is a region which itself contains numerous neurons with vocalization-correlated activity of a pre-onset type [21]. In contrast to the adjacent reticular formation, however, it lacks almost completely audio-sensitive cells [21]. The auditory input reaches the peri-collicular reticular formation mainly from the nucleus of the brachium of the inferior colliculus [20,24,25], with a minor contingent coming from the nucleus of the central acoustic tract [4]. Both these structures represent audiovocal interface structures themselves. In bats, their destruction causes a loss of Doppler shift compensation behavior, that is, the animals are no more able to compensate changes in the frequency of the echoes of their calls by changing the frequency of the emitted calls [28]. As normal bats adjust their calls only to their own echoes, but not to those of other animals, Doppler shift compensation, like the squirrel monkey’s duetting, depends upon the ability to discriminate selfproduced sounds from external sounds. Different reactions to self-produced and external sounds do not only exist on the brainstem level, but also on the ¨ cortical level. According to Muller-Preuss and Ploog [26], about 50% of the neurons in the secondary auditory cortex of the squirrel monkey, reacting to the play-back of vocalizations, show a weaker reaction or even no reaction at all to self-produced vocalization. The secondary auditory cortex receives a direct inhibitory input from the anterior cingulate cortex, another vocal control area apart from the periaqueductal gray [27]. This cortico-cortical connection also has been interpreted as a feedforward

164


mechanism by which a vocal motor structure controls auditory input. The question arises of how the cortical pathway differs from the brainstem pathway functionally. A possible difference could be that the anterior cingulate cortex provides the auditory cortex with information about the specific call type to be produced, while the periaqueductal gray informs the inferior colliculus only about the fact that there will be a vocalization, without specifying it. For a simple distinction of whether a perceived vocalization was self-produced or externally produced, it is not necessary to carry out an elaborate acoustic analysis; the feedforward information from the vocal control structure to the auditory structure that a phonation is started, would be sufficient. If, however, an external acoustic stimulus has to be analyzed which occurs during a self-produced vocalization, then specific knowledge about the acoustic characteristics of the self-produced vocalization is necessary for the analysis of the external stimulus. The present study does not exclude the possibility that there are also other mechanisms, apart from the vocalmotor feedforward mechanism, by which an organism can distinguish self-produced from foreign vocalizations. Proprioceptive feedback from the phonatory muscles or special acoustic features of the self-produced calls in comparison to those of others, might provide important cues as well. We still know very little about self-perception processes. The reticular formation bordering the inferior colliculus is, at least, a good candidate in this respect, deserving further investigation.

References [1] L.M. Aitkin, D.R.F. Irvine, W.R. Webster, Central neural mechanisms of hearing, in: I. DarianSmith (Ed.), Sensory Processes, Handbook of Physiology. Section 1. The Nervous System, Vol. III, American Physiological Society, Bethesda, MD, 1984, pp. 675–737. [2] L.M. Aitkin, W.R. Webster, J.L. Veale, D.D. Crosby, Inferior colliculus: I. Comparison of response properties of neurons in central, pericentral and external nuclei of adult cat, J. Neurophysiol. 38 (1975) 1196–1207. [3] E.E. Bauer, A. Klug, G.D. Pollak, Features of contralaterally evoked inhibition in the inferior colliculus, Hear. Res. 141 (2000) 80–96. [4] O. Behrend, G. Schuller, The central acoustic tract and audiovocal coupling in the horseshoe bat, Rhinolophus rouxi, Eur. J. Neurosci. 12 (2000) 4268–4280. [5] L.A. Benevento, L.B. McCleary, An immunohistochemical method for marking microelectrode tracks following single-unit recordings in long surviving, awake monkeys, J. Neurosci. Methods 41 (1992) 199–204. [6] A.A. Cameron, I.A. Khan, K.N. Westlund, K.D. Cliffer, W.D. Willis, The efferent projections of the periaqueductal gray in the rat: A Phaseolus vulgaris leucoagglutinin study, J. Comp. Neurol. 351 (1995) 568–601. [7] L.H. Carney, T.C.T. Yin, Responses to low-frequency cells in the inferior colliculus to interaural time differences of clicks: excitatory and inhibitory components, J. Neurophysiol. 62 (1989) 144–161. [8] J.H. Casseday, E. Covey, A neuroethological theory of the operation of the inferior colliculus, Brain Behav. Evol. 47 (1996) 311–336.

[9] J.R. Coleman, W. Clerici, Sources of projections to subdivisions of the inferior colliculus in the rat, J. Comp. Neurol. 262 (1987) 215–226. ¨ ¨ U. Hausler, ¨ ¨ [10] F. Dusterhoft, U. Jurgens, On the search for the vocal pattern generator. A single-unit recording study, Neuroreport 11 (2000) 2031–2034. [11] R. Emmers, K. Akert, A Stereotaxic Atlas of the Brain of the Squirrel Monkey (Saimiri sciureus), University of Wisconsin Press, Madison, WI, 1963. [12] K.A. FitzPatrick, Cellular architecture and topographic organization of the inferior colliculus of the squirrel monkey, J. Comp. Neurol. 164 (1975) 185–208. [13] R.F. Huffman, O.W. Henson Jr., The descending auditory pathway and acousticomotor systems: Connections with the inferior colliculus, Brain Res. Rev. 15 (1990) 295–323. ¨ [14] U. Jurgens, Vocalization as an emotional indicator. A neuroethological study in the squirrel monkey, Behaviour 69 (1979) 88–117. ¨ [15] U. Jurgens, The role of the periaqueductal gray in vocal behaviour, Behav. Brain Res. 62 (1994) 107–117. ¨ [16] U. Jurgens, Neural pathways underlying vocal control, Neurosci. Biobehav. Rev. 26 (2002) 235–258. ¨ [17] U. Jurgens, D. Ploog, On the motor coordination of monkey calls, in: J.D. Newman (Ed.), The Physiological Control of Mammalian Vocalizations, Plenum, New York, 1988, pp. 7–19. ¨ [18] U. Jurgens, R. Pratt, Role of the periaqueductal grey in vocal expression of emotion, Brain Res. 167 (1979) 367–378. ¨ [19] A. Kirzinger, U. Jurgens, The effects of brain stem lesions on vocalization in the squirrel monkey, Brain Res. 358 (1985) 150– 162. [20] M. Kudo, T. Tashiro, S. Higo, T. Matsuyama, S. Kawamura, Ascending projections from the nucleus of the brachium of the inferior colliculus in the cat, Exp. Brain Res. 54 (1984) 203–211. [21] C.R. Larson, M.K. Kistler, Periaqueductal gray neuronal activity associated with laryngeal EMG and vocalization in the awake monkey, Neurosci. Lett. 46 (1984) 261–266. [22] R.Y. Litovsky, T.C.T. Yin, Physiological studies of the precedence effect in the inferior colliculus of the cat. II. Neural mechanisms, J. Neurophysiol 80 (1998) 1302–1316. [23] P.W. Mantyh, Connections of midbrain periaqueductal gray in the monkey, J. Neurophysiol. 49 (1983) 567–594. [24] W. Metzner, Anatomical basis for audio-vocal integration in echolocating horseshoe bats, J. Comp. Neurol. 368 (1996) 252–269. [25] J.K. Moore, F. Karapas, R.Y. Moore, Projections of the inferior colliculus in insectivores and primates, Brain Behav. Evol. 14 (1977) 301–327. ¨ [26] P. Muller-Preuss, D. Ploog, Inhibition of auditory cortical neurons during phonation, Brain Res. 215 (1981) 61–76. ¨ ¨ [27] P. Muller-Preuss, J.D. Newman, U. Jurgens, Anatomical and physiological evidence for a relationship between the cingular vocalization area and the auditory cortex in the squirrel monkey, Brain Res. 202 (1980) 307–315. [28] INABIS 98, CD-ROM G. Schuller, Neural mechanisms of vocal control in bats, in: S. Brudzynski (Ed.), Neural Mechanism of Mammalian Vocalization, McMaster University, Canada, 1999. [29] G. Schuller, S. Fischer, H. Schweizer, Significance of the paralemniscal tegmental area for audio-motor control in the mustached bat, Pteronotus p. parnellii: The afferent and efferent connections of the paralemniscal area, Eur. J. Neurosci. 9 (1997) 342–355. [30] A. Tokunaga, S. Sugita, K. Otani, Auditory and non-auditory subcortical afferents to the inferior colliculus in the rat, J. Hirnforsch. 25 (1984) 461–472. [31] M. Wiberg, J. Westman, A. Blomqvist, Somatosensory projection to the mesencephalon: an anatomical study in the monkey, J. Comp. Neurol. 264 (1987) 92–117. ¨ ¨ [32] A. Wienicke, U. Hausler, U. Jurgens, Auditory frequency discrimination in the squirrel monkey, J. Comp. Physiol. A 187 (2001) 189–195.

Research report N euronal activity in the inferior

Recommend Documents