Biological Cybemetics
Biol. Cybern. 67, 501-509 (1992)
9 Springer-Verlag 1992
Dynamic characteristics of the auditory cortex of Guinea pigs observed with multichannel optical recording Kohyu Fukunishi, Nobuyuki Murai, and Hiroyuki Uno Advanced Research Laboratory, Hitachi Ltd., Hatoyama, Saitama 350-03, Japan Received January 7, 1992/Accepted in revised form May 20, 1992
Abstract. The spatiotemporal characteristics of neural activity in the guinea pig auditory cortex are investigated to determine their importance in neural processing and coding of the complex sounds. A multi-channel optical recording system has been developed for observing the cortical field of the mammalian brain in vivo. Using the voltage-sensitive dye: RH795, optical imaging was used to visualize neural activity in the guinea pig auditory cortex. Experimental results reveal a boomerang-shaped pattern of movement of activated neural cell regions for the evoked response to click as complex sounds. Parallel and sequential neural processing structure was observed. Although the exact frequency selectivities of single cells and tonotopical organization observed using microelectrode were not visible, the similar feature to the microelectrode evidences was imaged by extracting the strongly response field from the optical data.
Introduction The general functions of the auditory cortex in the mammalian brain are as yet not well understood. Many relevant hypotheses are discussed by Pickles (1988). The main function of the auditory cortex is the analysis of complex sounds to filter for vocal sounds or other species-specific sounds. In studying the principle of and the comprehension mechanism for representing, and analyzing complex sounds, many efforts have been devoted to various aspects of the topographic representation of complex signals in the auditory nervous system from the inner ear to the auditory cortex, discussed in the proceedings of the meetings edited by Syka and Aitkin (1981) and Edelman et al. (1988). The principal characteristic of the topographic representation is the tonotopic arrangement of cells (the tonotopic organization) in the various nuclei in the auditory system, which Correspondence to: K. Fukunishi
indicates the ear's frequency selectivity and frequency analyzing ability. This tonotopic organization is maintained in the mammalian auditory cortex of many different species, such as the cat, monkey, squirrel and guinea pig (Merzenich et al. 1988 for review). The functional role of tonotopical organization in the analysis of complex sounds in the auditory cortex, however, remains unclear, despite efforts to describe tonotopical mapping through microelectrode studies (Hellweg et al. 1977; Redies et al. 1989; Irvine and Robertson 1990). A simple question remains: why is the tonotopic organization found in every nuclei of the auditory pathway maintained in the .auditory cortex? The acoustic signal disappears in an instant after having been emitted, even while the cognitive processing for the signal is performed in the auditory processing brain. Therefore, frequency selectivity in the cortex would make the higher-order processing of the complex sounds intolerably inefficient. Hitherto, recording of the electrical events in single auditory neural cells and fibers has been used to investigate topographic mapping in the auditory system, including the cortex. However, complex sound mapping in the auditory cortex is riot adequately depicted by the activity of individual cells except for the acoustical specific animal as the bat and the barn owl. A group of cells seems to work together for such mapping. The temporal relations of the neural activity among the groups of cells distributed in the auditory cortex suggest that there is a special mechanism such as sequential and distributed parallel processing for the neural processing of the complex sounds. Tonotopic mapping represents only the spatial information of the auditory neural cells in the cortex and does not describe any temporal information of the neural activity. On the other hand, the optical imaging method using voltage-sensitive dyes, introduced by Cohen et al. (1974) has proven successful for observing spatiotemporal neural activity in mammal brains as described in books edited by De Weer and Salzberg (1986) and Loew (1988) and as reviewed in detail by Grinvald et
502 A separate optical imaging technique without dyes measures intrinsic signals related to neural activity. This technique has been employed to record images of the ocular dominance organization in the living monkey striate cortex - V1 - (D. Y. Ts'o et al. 1990). Although optical imaging of intrinsic signals might avoid the disadvantages of pharmacological side effects, phototoxicity and staining uncertainly as seen in the optical imaging using dyes, the signal level was very low, only one to ten percent of the signal level when using dyes. Therefore, only pseudo-static neuron activity was able to be imaged with a low frame frequency CCD camera and the optical signals for 30 min were averaged to produce the orientation image (D. Y. Ts'o et al. 1990). These results with optical imaging of the cortical field have been reported for the visual cortex, no such reports have been published on the auditory cortex. This work reports some results of preliminary experiments on imaging the spatiotemporal characteristics of neural activities and the topographical mapping of acoustic signals in the auditory cortex of the guinea pig. Experiment used voltage-sensitive dye and speciallydesigned optical recording system with a 12 x 12 photodiode array. The current work follows initial experiment executed using a small-scale prototype optical imaging system with a 2 • 2 photodiode array (Fukunishi et al. 1989).
al. (1988). This method permits simultaneous observation of adjacent individual neural cell (neuron) activity after the stimulus by monitoring with a silicon photodiode array or CCD sensor. The voltage-sensitive dyes translate the electrical characteristic of potential response, associated with the spikes of individual neuron activity, into an optical characteristic. A review by Grinvald (1985) covers tests of several hundred dyes to determine the optimum dye for the optical imaging, one that produces a sufficiently large optical signal, which changes synchronously with the spike activity and that introduces less pharmacological side effects. This optical imaging method has been used to study the topographic organization of the neurons and the interrelationships of the neural activity in the receptive field of mammal cortices in vivo. The first proof of the method's usefulness was in Orbach and Cohen's (1983) observation o f electrically-evoked responses in the salamander olfactory bulb. Later, this approach was used to observe naturally-evoked neuron activity in intact frog brains as they responded to a light pattern displayed on an oscilloscope screen. Grinvald et al. (1984) discussed the potentiality for imaging the retinotectal connection in the visual receptive field in the tectum using this method. Similarly, Orbach et al. (1985) recorded and imaged the evoked response pattern of whisker barrel activity in the rat somatosensory system resulting from a whisker movement. Furthermore, an important effort using voltage-sensitive dyes and video image processing has resolved the long-standing question of cortical organization in relation to ocular dominance and orientation columns of the striate field of the monkey visual cortex (Blasdel and Salama 1986). Research continues on column orientation in the visual cortices of cats and monkeys thanks to improvement in the technical aspects of optical imaging.
I/V '.... -'~1::>
CCD
Fluorescence optical measurement was used in the optical imaging of the parallel processing structure of the auditory cortex. The apparatus used in the experiments is shown schematically in Fig. 1. The optical system was specially designed to maintain high transparency, that is, to prevent signal loss in the optical path, and to keep
[>
12xl 2 diode array . ~ "
j~j
Biol. Cybern. 67, 501-509 (1992)
9 Springer-Verlag 1992
Dynamic characteristics of the auditory cortex of Guinea pigs observed with multichannel optical recording Kohyu Fukunishi, Nobuyuki Murai, and Hiroyuki Uno Advanced Research Laboratory, Hitachi Ltd., Hatoyama, Saitama 350-03, Japan Received January 7, 1992/Accepted in revised form May 20, 1992
Abstract. The spatiotemporal characteristics of neural activity in the guinea pig auditory cortex are investigated to determine their importance in neural processing and coding of the complex sounds. A multi-channel optical recording system has been developed for observing the cortical field of the mammalian brain in vivo. Using the voltage-sensitive dye: RH795, optical imaging was used to visualize neural activity in the guinea pig auditory cortex. Experimental results reveal a boomerang-shaped pattern of movement of activated neural cell regions for the evoked response to click as complex sounds. Parallel and sequential neural processing structure was observed. Although the exact frequency selectivities of single cells and tonotopical organization observed using microelectrode were not visible, the similar feature to the microelectrode evidences was imaged by extracting the strongly response field from the optical data.
Introduction The general functions of the auditory cortex in the mammalian brain are as yet not well understood. Many relevant hypotheses are discussed by Pickles (1988). The main function of the auditory cortex is the analysis of complex sounds to filter for vocal sounds or other species-specific sounds. In studying the principle of and the comprehension mechanism for representing, and analyzing complex sounds, many efforts have been devoted to various aspects of the topographic representation of complex signals in the auditory nervous system from the inner ear to the auditory cortex, discussed in the proceedings of the meetings edited by Syka and Aitkin (1981) and Edelman et al. (1988). The principal characteristic of the topographic representation is the tonotopic arrangement of cells (the tonotopic organization) in the various nuclei in the auditory system, which Correspondence to: K. Fukunishi
indicates the ear's frequency selectivity and frequency analyzing ability. This tonotopic organization is maintained in the mammalian auditory cortex of many different species, such as the cat, monkey, squirrel and guinea pig (Merzenich et al. 1988 for review). The functional role of tonotopical organization in the analysis of complex sounds in the auditory cortex, however, remains unclear, despite efforts to describe tonotopical mapping through microelectrode studies (Hellweg et al. 1977; Redies et al. 1989; Irvine and Robertson 1990). A simple question remains: why is the tonotopic organization found in every nuclei of the auditory pathway maintained in the .auditory cortex? The acoustic signal disappears in an instant after having been emitted, even while the cognitive processing for the signal is performed in the auditory processing brain. Therefore, frequency selectivity in the cortex would make the higher-order processing of the complex sounds intolerably inefficient. Hitherto, recording of the electrical events in single auditory neural cells and fibers has been used to investigate topographic mapping in the auditory system, including the cortex. However, complex sound mapping in the auditory cortex is riot adequately depicted by the activity of individual cells except for the acoustical specific animal as the bat and the barn owl. A group of cells seems to work together for such mapping. The temporal relations of the neural activity among the groups of cells distributed in the auditory cortex suggest that there is a special mechanism such as sequential and distributed parallel processing for the neural processing of the complex sounds. Tonotopic mapping represents only the spatial information of the auditory neural cells in the cortex and does not describe any temporal information of the neural activity. On the other hand, the optical imaging method using voltage-sensitive dyes, introduced by Cohen et al. (1974) has proven successful for observing spatiotemporal neural activity in mammal brains as described in books edited by De Weer and Salzberg (1986) and Loew (1988) and as reviewed in detail by Grinvald et
502 A separate optical imaging technique without dyes measures intrinsic signals related to neural activity. This technique has been employed to record images of the ocular dominance organization in the living monkey striate cortex - V1 - (D. Y. Ts'o et al. 1990). Although optical imaging of intrinsic signals might avoid the disadvantages of pharmacological side effects, phototoxicity and staining uncertainly as seen in the optical imaging using dyes, the signal level was very low, only one to ten percent of the signal level when using dyes. Therefore, only pseudo-static neuron activity was able to be imaged with a low frame frequency CCD camera and the optical signals for 30 min were averaged to produce the orientation image (D. Y. Ts'o et al. 1990). These results with optical imaging of the cortical field have been reported for the visual cortex, no such reports have been published on the auditory cortex. This work reports some results of preliminary experiments on imaging the spatiotemporal characteristics of neural activities and the topographical mapping of acoustic signals in the auditory cortex of the guinea pig. Experiment used voltage-sensitive dye and speciallydesigned optical recording system with a 12 x 12 photodiode array. The current work follows initial experiment executed using a small-scale prototype optical imaging system with a 2 • 2 photodiode array (Fukunishi et al. 1989).
al. (1988). This method permits simultaneous observation of adjacent individual neural cell (neuron) activity after the stimulus by monitoring with a silicon photodiode array or CCD sensor. The voltage-sensitive dyes translate the electrical characteristic of potential response, associated with the spikes of individual neuron activity, into an optical characteristic. A review by Grinvald (1985) covers tests of several hundred dyes to determine the optimum dye for the optical imaging, one that produces a sufficiently large optical signal, which changes synchronously with the spike activity and that introduces less pharmacological side effects. This optical imaging method has been used to study the topographic organization of the neurons and the interrelationships of the neural activity in the receptive field of mammal cortices in vivo. The first proof of the method's usefulness was in Orbach and Cohen's (1983) observation o f electrically-evoked responses in the salamander olfactory bulb. Later, this approach was used to observe naturally-evoked neuron activity in intact frog brains as they responded to a light pattern displayed on an oscilloscope screen. Grinvald et al. (1984) discussed the potentiality for imaging the retinotectal connection in the visual receptive field in the tectum using this method. Similarly, Orbach et al. (1985) recorded and imaged the evoked response pattern of whisker barrel activity in the rat somatosensory system resulting from a whisker movement. Furthermore, an important effort using voltage-sensitive dyes and video image processing has resolved the long-standing question of cortical organization in relation to ocular dominance and orientation columns of the striate field of the monkey visual cortex (Blasdel and Salama 1986). Research continues on column orientation in the visual cortices of cats and monkeys thanks to improvement in the technical aspects of optical imaging.
I/V '.... -'~1::>
CCD
Fluorescence optical measurement was used in the optical imaging of the parallel processing structure of the auditory cortex. The apparatus used in the experiments is shown schematically in Fig. 1. The optical system was specially designed to maintain high transparency, that is, to prevent signal loss in the optical path, and to keep
[>
12xl 2 diode array . ~ "
j~j
