Takuji Kasamatsu
Senior Scientist
Smith-Kettlewell Eye Research Institute
2318 Fillmore Street
San Fransicso, CA 94115
Phone: 415-345-2000 (receptionist)
415-345-2173 (office)
415-345-2170 (lab)
415-345-8455 (FAX)
E-mail: takuji@ski.org
Neuroscience research in my laboratory has two directions: one is the functional organization of mammalian visual cortex and the other, neuronal plasticity typically seen in developing visual cells of the thalamus and neocortex. The latter topic has a long history, while the former has received more focussed attention recently.
Collinearity in Perceptual Grouping
How we perceive an object in the visual field is an intriguing question in visual neuroscience. Gestalt perception theory predicts that the visual system groups features together based on similarity, proximity, smoothness or closure. Recent psychophysical and neurophysiological studies have found evidence of spatial facilitation and suppression that appears to conform to these "perceptual grouping" rules. These contextual grouping procedures may be implemented by long-range lateral interactions in the visual cortex.
For the last few years my laboratory has studied the neural basis of this matter, recording single-cell firing from the striate cortex of anesthetized and paralyzed preparations. We use this preparation rather than behaving animals, since it is free from potential complications associated with "attention". The spatial profile of the CRF, especially that of simple cells, is well-fitted by a Gabor pattern (a sinusoidally varying luminance distribution weighted by a one-dimensional Gaussian envelope). Therefore, we use Gabor patches as the stimuli in our study. We found that single-cell responses to a central target were modulated by collinear flanking targets. This modulation was further dependent on the cell's contrast threshold for the central target: many cells showed facilitation at near-threshold contrast, but suppression at supra-threshold contrast (Fig.1. See also Mizobe et al., ARVO abstr. 37, S483, 1996).
Based on these preliminary findings we have hypothesized that the lateral interactions in the striate cortex are arranged into two antagonistic mechanisms: 1) facilitation that is spatially organized along the optimal orientation of the cell in a collinear fashion, and 2) suppression that is less-selective for orientation or spatial frequency and that is distributed diffusely around the cell's classical receptive field (CRF). The balance between the two mechanisms, dictated by the cell's contrast threshold, seems to control the cell's firing behavior. We are currently testing this hypothesis using various contrasts and configurations of the stimulus pattern in combination.
Retinotopic and Non-retinotopic Responses
Earlier, we have sought physiological correlates of long-range lateral connections through axon collaterals of pyramidal cells in the striate cortex, using a method different from that described above. We devised a physiological method to characterize short- and long-range lateral connections functionally. We differentially recorded local field potentials using a pair of microelectrodes placed in the striate cortex near the projection site of the visual field center (i.e., center of gaze). A particular strength of our method is that we can stimulate many loci in the visual field concurrently, identifying individual neural responses elicited by each of many stimuli. Thus, we can describe in detail the dynamics of neural interactions among them, for example, keeping one stimulus at the CRF and others at various places in its surround. This procedure was made possible by using a nonlinear systems analysis method developed by E.E. Sutter at our institute. A generic version of the software, VERIS TM, is widely used in various clinical applications, including multiple electro-retinogram recording.
We found that local field potentials comprised two components. One is a short-latency, fast-local component (FLC) with a retinotopic organization, similar to that seen with single-cell discharges elicited by stimulation of the CRF at the same site. The other is a slow non-retinopotic component with a long peak latency. The slow-distributed component (SDC) had an extensive receptive field. There were strong inhibitory interactions between the FLC and SDC, and among the SDCs elicited by stimulation of different loci outside the CRF in the visual field (Kitano et al., Visual Neuroscience, 11, 953-977, 1994). The FLC and SDC are thought to be a physiological correlate of the short- and long-range lateral connections, respectively.
Differential control by GABAA and GABAB
We further noted that the two components are regulated by two types of GABA receptors differentially. The FLC is more strongly suppressed by a GABAA receptor agonist (muscimol) than GABAB agonist (baclofen) and the reverse is the case for the SDC (Mizobe et al., Abstr. Soc. Neurosci.., 21, 1655, 1995). GABAB receptors are known to be localized mostly at distal dendrites (and presynaptic fibers), though GABAA receptors are found throughout the neuronal membrane. Taken together with this different distribution pattern of the two types of GABA receptors, the finding seems to suggest that the long-range axon-collateral projection for the SDC ends on the distal dendrites and that the short-range fiber projection for the FLC on the proximal dendrites closer to the soma.
Lack of Lateral Interactions in Monocular Deprivation
More recently, we have obtained evidence that the long-range lateral interactions are missing from the visual cortex of amblyopic animals. The normal FLC and SDC were elicited by stimulation of the normal, untreated eye. However, stimulating the monocularly lid-sutured eye, we failed to elicit the normal potentials. Instead, we obtained abnormal responses with long latencies by stimulation of the retinotopic locus of the recording site. No SDCs were elicited by remote stimulation, either. Thus, the monocularly deprived cortex lacked the spatial interactions which were characteristic of the normal visual cortex (Kasamatsu et al., in preparation).
Neurochemical Changes Underlying Amblyopia
The above study on monocularly deprived animals leads us to another main line of research in my laboratory which started nearly 20 years ago and which is still pursued on a reduced scale. This is study on an animal model of the neurochemical basis of lazy eye syndromes (i.e., amblyopia).
Development and Experience
Babies of some mammals including humans are, arguably, prematurely born fetuses. Their brain matures rapidly during the first few weeks or months of life under parental care outside the uterus. Throughout this period of protracted "neurogenesis" after birth, the central visual system remains susceptible to the precise physical nature of the environment. That is, provided that the genetic contribution is normal, "visual experience" received by the developing visual center shapes a perfectly normal visual system in the brain(Fig.2).
However, the outcome of this sensory-motor experience in early life is not always beneficial; sometimes adverse circumstances such as severe sensory deprivation, crossed eyes and high refractive error do not allow the visual brain to develop normally and the individual suffers from perma nent visual loss (amblyopia). The neural mechanisms underlying this age-dependent and experience-regulated plasticity or modifiability of nerve-cell connections in the developing brain are largely unknown .
Ocular Dominance Plasticity
Most cells in normal visual cortex are binocular and respond with various degrees of strength to visual stimuli delivered to the two eyes (ocular dominance). A useful animal model of experience regulated cortical plasticity is to study changes in ocular dominance following a period of eyelid suture of one eye, while animals are still young.
Fig. 2
The monocularly deprived eye becomes functionally "blind" due to loss of its connections with the visual center and most cells respond only to stimulation of the non-deprived eye.
The Noradrenaline Hypothesis
Some years ago, we proposed that the noradrenaline (NA)-activated, beta-adrenoreceptor system in the brain is necessary for maintaining a high level of "ocular dominance plasticity" in the visual center of developing animals. [Fig. 3. For a critical review, see Kasamatsu, T. Studies on regulation of ocular dominance plkasticity: strategies and findings (1994) Exp. Brain Res. SR 24, 68-80]
Fig. 3
For the last 33 years, this type of neural plasticity has been customarily thought to be present only in the visual center of visually "attentive" animals in action. However, using a technique of continuous infusion of NA directly into the visual center, we have established that essentially the same plasticity is working in the visual center of anesthetized and paralyzed animals. [For a critical review, see Imamura K. and Kasamatsu T., Ocular dominance plasticity: usefulness of anesthetized and paralysed preparations. (1991) Jap. J. Physiol. 41, 521-549] This result indicates that the NA-beta adrenoreceptor system is indeed a neurochemical basis of neural plasticity in the develop ing visual center. By applying the same methodology, we have also shown the partial restoration of neural plasticity to the visual center of adult animals that under usual conditions are mostly insensitive to environmental influences. We are further pursuing neurochemical mechanisms of age-dependent, experience-regulated neural plasticity. Our long-term goal is to provide neurochemical bases for drug-assisted orthoptics or visual rehabilitation.
Collaborators:
