| |
| Masashi Kawasaki |
| General |
|
A major function of the central nervous system is to recognize complex patterns of stimulus parameters in order to appropriately control behavior. Many forms of animal behavior are not merely responses to simple physical parameters such as intensity of light, frequency of sound, or temperature; rather, behaviors are often controlled by a complex pattern of stimulus parameters in both space and time, and the information required to generate a correct behavioral output is often hidden in that pattern. The task of the central nervous system is to extract such hidden information by means of neuronal computation and send it to the motor system to generate appropriate behavior. The long-term goal of my research career is to track the flow of information for natural behaviors in a vertebrate system from sensory receptor neurons through the central nervous system and finally to the motor output. My work focuses on the integral aspects of information processing in the brain of fish, whose central nervous system is relatively simple among vertebrates yet which shares organizational principles with more advanced systems, such as vision, audition and motor systems, in higher vertebrates. I am also pursuing the evolutionary understanding of neuronal circuits. A nervous system presents us with two fundamentally different questions: how does it work and why does it work that way? While physiological studies in a given species will answer the former question, only comparative studies using different species can answer the latter. Two unrelated species of electric fishes, Eigenmannia and Gymnarchus, provide a good model system to address this evolutionary issue (see below for more detail). I am further interested in the comparative behavioral physiology of South American and African species of electric fishes. These electric fishes provide a rare opportunity, in which comparison of their behavioral, sensory and motor functions is possible in a phylogenetic frame work. |
| Current and future research focus: |
|
Since becoming an assistant professor at the University of Virginia, I have been working on an African electric fish, Gymnarchus, which has evolved electrosensory systems independently from the South American electric fish, Eigenmannia, that I extensively studied in Heiligenberg's laboratory (1986-1990). Despite their independent evolution, these two taxa perform very similar jamming avoidance responses. After conquering a major technical difficulty, I have identified complex computational algorithms for the JAR in Gymnarchus and discovered that these fishes use a series of identical computational methods. I found that, in striking contrast to closely related mormyrids, Gymnarchus does not use an internal reference mechanism to recognize its own electric organ discharge(21). Comparative physiology and anatomy of the pacemaker further shows that closely related fishes, Gymnarchus and mormyrid electric fishes, possess similarly organized yet functionally different pacemakers. Moreover, the pacemaker of Gymnarchus is functionally similar to that of unrelated Eigenmannia.(24) I am currently working on the sensory mechanisms for the jamming avoidance response in Gymnarchus*specifically, the mechanisms for phase comparison, an essential computational feature for the behavior, in the electrosensory lateral line lobe. We are discovering that the same computational task, phase comparison, is performed by different brain structures in Gymnarchus and Eigenmannia. We recently characterized the phase comparison circuit in Gymnarchus physiologically and anatomically(25) and found that neurons in the electrosensory lateral line lobe that perform phase comparisons are sensitive to a few microseconds of phase difference. Accordingly, the phase-locked neurons which provide input to the differential phase comparator neurons are high regular (standard deviation of jitter ~ 5msec)31. I plan to investigate the ultimate cellular mechanisms for this phase accuracy using intracellular recording and labeling techniques. Currently, we are successfully recording from and labeling the phase comparator neurons with the in vivo whole-cell recording technique. These studies are supported by an Research Scientists Development Awards grant from the NIMH (1995-2000, $353,000) and a grant from NSF (1996-1999, $ 271,200). In the coming several years, I plan to perform physiological and anatomical studies on Gymnarchus towards the long-term goal of elucidating the entire neuronal mechanism for the jamming avoidance response, from sensory receptors to behavioral output. There are many topics at different levels, each of which has its own intrinsic interests and requires different skills. They are specifically, (1) Parelell processing of amplitude and phase information(32), (2) Resolution of spatial ambiguity by central neurons, (3) Generation of motor codes by premotor neurons, (4) Temporal hyperacuity at the behavioral level(22,28). By comparing results from Gymnarchus with those already obtained from South American fishes by Heiligenberg and his colleagues, I will attempt to identify and separate general principles and specific solutions among the features of neuronal computation in these systems. This effort will be a rare opportunity to compare how complex neuronal circuitry evolved independently to perform the same task, and to address the question why the nervous system works the way it works(21,27,30). An additional work was published in collaboration with Walter Heiligenberg since I came to the University of Virginia(19). A behavioral experiment I performed on Gymnarchus at University of Virginia(21) cast doubt upon the interpretation of some behavioral studies by Heiligenberg. We collaborated on a theoretical work which demonstrated that the original theory for the jamming avoidance response was correct and explained more accurately why fish can perform 'correct' jamming avoidance responses even when a very large jamming signal was applied. The third line of study is on the frequency falling behavior in a pulse-type gymnotiform fish, Rhamphichthys. We recently found through behavioral experiments that this fish uses different computational rules for its slow frequency shifting behaviors than those used by wave-type gymnotiform electric fishes. This study was carried out with an undergraduate student, Jon Prather, and has been published(26). I have recently added two research topics in Eigenmannia - detection of frequency modulation by the ampullary electrosensory system(33), frequency decreasing behavior in response to pure amplitude modulation(34). |
| Scientific background and past research accomplishment before University of Virginia |
|
My serious interest in animal behavior began when I read a series of books by classical ethologists such as Konrad Lorenz and Niko Tinbergen in college. This literature struck me with its outstanding view of animal behavior as an information processing system. Since then, my interests have focused on information processing by the central nervous system. While I am continuously learning modern neurobiological concepts and techniques to understand elementary processes in the system, such as membrane and synaptic functions and cell-to-cell signal communication, I always try to understand the neuronal code that those elementary processes carry for information processing. This focus will remain for my entire research career. For my dissertation work at Sophia University in Tokyo, I studied central visual neurons in the Japanese dace (cyprinid fish), which exhibits very rapid swimming behavior, and I found that neurons in the optic tectum encode very rapid movements of visual images1,3. As a part of my dissertation study under the guidance of Ken-Ichi Naka, I also examined dynamic properties of the retinal neurons of catfish (Ictalurus) using a Wiener-kernel cross-correlation technique. I found that visual sensitivity to steady state signals and dynamic signals in the horizontal cells of catfish could be separately regulated(2). As a postdoc in Nobuo Suga's laboratory at Washington University in St. Louis (1984-1986), I studied a more complex system, the echolocation system of bats. In collaboration with Prof. Nobuo Suga and Dr. Daniel Margoliash, I demonstrated that neurons in the auditory cortex of the bat are sensitive to a combination of vocal self-stimulation due to the bat's own cry and to its echo. These two stimuli together cause a strong non?linear summation that codes an important stimulus feature, the presence of a target. Moreover, these neurons are tuned to a particular echo delay which encodes a corresponding target distance5. In Prof. Suga's laboratory, I was also involved in a work in which I characterized a higher?order cortical map, measured the directional selectivity of the FM-FM combination sensitive neurons, and estimated the intensity of vocal self-stimulation(6,7). These studies dealt with specific physiological stages for information processing. I became more interested in a single preparation in which I could examine large-scale information flow for behavioral output. I chose electric fish. In 1986, I moved to Prof. Walter Heiligenberg's laboratory at University of California at San Diego. In collaboration with Prof. Walter Heiligenberg and Dr. Gary Rose, I identified physiologically and anatomically a diencephalic nucleus which is located at the top of a sensory hierarchy and also controls a specific behavior, the jamming avoidance response (JAR). Neurons in this prepacemaker nucleus respond specifically to a complex spatio?temporal pattern of the stimulus that controls the JAR, and they unambiguously encode a critical feature of it(11). Stimulation of these neurons elicits behavioral responses which are indistinguishable from their naturally occurring form(10). Identifying this nodal point at which a sensory hierarchy meets a motor command system enabled us to trace almost completely the signal pathways for a complex vertebrate behavior. These prepacemaker neurons are also significant in light of their temporal hyperacuity. The JAR displays an extraordinary sensitivity to small temporal disparities (as small as 400 nanoseconds). Together with Rose and Heiligenberg, I found that the prepacemaker neurons are sensitive to temporal disparities as small as 1 microsecond8. Because individual electroreceptors are incapable of encoding such small timing, the behavioral sensitivity must necessarily be an emergent property of the nervous system. I found in a recent study in Gymnarchus that the very accurate EOD is not necessary for the hyperacuity behavior(31). The prepacemaker has yet another interesting aspect. We have found that this nucleus is subdivided into two regions which each control distinctly different behaviors*the JAR and sexual and agonistic responses. I have found that the intracellular stimulation of a single prepacemaker neuron could cause 'chirps', a rapid pacemaker frequency modulation shown in sexual and agonistic encounters(9,10). At the axonal terminals of prepacemaker neurons in the pacemaker nucleus, different glutamate receptor subtypes mediate different behaviors(12,16). I also studied sexual behavior of another electric fish, Hypopomus brevirostris, recording six distinct patterns of electric organ discharges during courtship and spawning. I subsequently induced all of these behaviors independently by stimulating different parts of the diencephalon in an immobilized preparation. This system allowed me to investigate cellular mechanisms in the pacemaker nucleus which are responsible for specific behavioral patterns15. This ongoing study is supported by an R29 grant from the National Institute of Mental Health (1992-1998, $350,000). These studies on electric fish at University of California at San Diego formed a strong basis for my current research conceptually and technically and are contributing to the closely related studies currently pursued at the University of Virginia. |
| References: |
|