Hippocampal formation, as a part of cerebral archicortex, is, said to be the site of “engrams” from studies on leison and stimulating experiment.
In this paper, a dynamic neural model of hippocampal CA 3-dentate area, which accounts for its physiological properties, has been developed to study hippocampal memory function. This model consists of several granule cells, pyramidal cells and basket cells, and it represents potential conduction within each cell. For this modeling two hypotheses are proposed ; (i) potentiate time constant at dendritic initial segment decreases temporarily according to potentiation in this segment, (ii) synaptic conductivity is reinforced or weakened according to the principle of effective or ineffective use of the synapse.
Computer simulation of physiological experiments supports these two hypotheses for “frequency potentiation” and “long lasting potentiation”, which are neural plasticity characteristically found in hippocampus.
When a spatial pattern with Poisson pulse is put in through the perforant path, the output pattern is determined by decrease of initial segment time constant and lateral inhibition by basket cells. This is the spatial pattern of pulse output from cells whose input frequency is maximum within inhibition area of a basket cell. After a few patterns are put in repeatedly and if an imperfect pattern of one of the previous pattern is put in, the output pattern becomes that of previous one. This means that the hippocampal model has an associative memory function, which is accomplished by synaptic conductivity change similar to correlation matrix memory models. From this result, hippocampus is considered to be a memory device for entorhinal area.

抄録全体を表示

This paper describes an automatic volumetry system which evaluates the volume of gallbladder using ultrasonic tomography. The volume of a three-dimensional object with smooth shape can be evaluated from its equally spaced consecutive cross sections. This system processes N (=5-8) equally spaced sectional images obtained by an ultrasonic diagnostic instrument and extracts N cross sectional regions of gallbladder automatically.
Region extraction algorithm used in this system is stable against noise which is inherent to ultrasonic images and adaptive to the gray-level changes. And it is time-effective and hardware oriented since it includes no multiplication operations. This system also has a function of self-diagnosis, which refuses to accept any extracted region which is far from what it should be.
Some results of experiments to show the effectiveness of this system are shown. It enables clinicians to measure the dynamics of the volume change of gallbladder, which will be very useful clinically.

抄録全体を表示

Parametric utilities based on local inputs and computations, named “PUBLIC”, a microcomputer DBMS, is introduced, as its medical application has been evaluated.
The “PUBLIC” has five functions including some data entry commands, data manipulation commands, data manipulation subroutines, statistical commands and data transmission commands. These functions are designed to be easily used by any person.
It has been concluded that “PUBLIC” was very useful for medical applications

抄録全体を表示