Takayuki Kannon

2011 Special Issue: PLATO: Data-oriented approach to collaborative large-scale brain system modeling

The brain is a complex information processing system, which can be divided into sub-systems, such as the sensory organs, functional areas in the cortex, and motor control systems. In this sense, most of the mathematical models developed in the field of neuroscience have mainly targeted a specific sub-system. In order to understand the details of the brain as a whole, such sub-system models need to be integrated toward the development of a neurophysiologically plausible large-scale system model. In the present work, we propose a model integration library where models can be connected by means of a common data format. Here, the common data format should be portable so that models written in any programming language, computer architecture, and operating system can be connected. Moreover, the library should be simple so that models can be adapted to use the common data format without requiring any detailed knowledge on its use. Using this library, we have successfully connected existing models reproducing certain features of the visual system, toward the development of a large-scale visual system model. This library will enable users to reuse and integrate existing and newly developed models toward the development and simulation of a large-scale brain system model. The resulting model can also be executed on high performance computers using Message Passing Interface (MPI).

A Next Generation Modeling Environment PLATO: Platform for Collaborative Brain System Modeling

To understand the details of brain function, a large scale system model that reflects anatomical and neurophysiological characteristics needs to be implemented. Though numerous computational models of different brain areas have been proposed, these integration for the development of a large scale model have not yet been accomplished because these models were described by different programming languages, and mostly because they used different data formats. This paper introduces a platform for a collaborative brain system modeling (PLATO) where one can construct computational models using several programming languages and connect them at the I/O level with a common data format. As an example, a whole visual system model including eye movement, eye optics, retinal network and visual cortex is being developed. Preliminary results demonstrate that the integrated model successfully simulates the signal processing flow at the different stages of visual system.

Generating realistic retinal image for whole visual system modeling

Recent progress and improvements on optical technology have enabled us to measure the characteristics of the eye in great detail. It has also been shown that, even when myopia and astigmatism are completely corrected with spectacle, the retinal image is still blurred due to uneven refraction power of the cornea and the crystalline lens [1][2], an effect known as “irregular astigmatism”. However, almost all researches in visual science assume sharp images as input. Currently, a large-scale whole visual system models is being developed on a supercomputer to elucidate its complex function [3]. Moreover, it has been demonstrated that the blurred retinal images are gradually compensated at both, the retina and the cortex [4]. Therefore, we have developed an eye optics model which can calculate the blurred retinal image based on known physiological evidences to provide a more realistic input to retinal and visual cortex model. The proposed model is described by the linear image filter based on Artals model [5]. It can process multi-spectral image for reproducing the spectral characteristics of photoreceptors. OTF (Optical transfer function) calculated from the wavefront aberration with spectral transmittance of the lens is used for image processing filter. The wavefront aberration is defined by Zernike Coefficients values (measured by wavefront aberrometer such as OPD-Scan or Shack-Hartman Sensor systems) or SCA values (power of sphero-cylindrical lens with its axis for spectacle lens or contact lens prescriptions). To consider the effects of diffraction, OTF is designed to vary depending on pupil diameter and wavelength. To account for aging effects, the spectral transmittance of the lens [6] was also implemented. As inputs of the model, multispectral images taken by multispectral camera or converted from RGB image can be utilized. We simulated a closed-loop system connecting the proposed model to a simple eye movement model and pupil model. The eye movement model generates random saccades and the pupil model modifies pupil diameter according to the mean luminance of retinal image. As a result, our model could reproduce the effect of chromatic aberration, pupil diameter and age on the blurred retinal image. By applying our model to the large-scale visual system model, it is expected to be able to evaluate the phenomena, for example, age dependent perception on visual illusions, blur compensation in retina or visual cortex, and so on.

Simulation platform: cloud-computing meets computational neuroscience

Computational models and theoretical tools are essential components in computational neuroscience. A number of models and tools have been developed and registered at various online databases such as ModelDB and J-Node Platforms. Yet, the reuse of such resources still remains quite difficult. For example, to carry out a computer simulation of a model, we have to download the program from the database, extract, read instructions, compile if the program is written in a general programming language such as C, install the appropriate neural simulator if it is written for a simulator such as GENESIS, NEURON, and NEST, and finally we may be ready to do it, if no problems occurs during all the setup mentioned above. How can we avoid this hustle? As a solution of it, we introduce a cloud-based system for online computer simulation called Simulation Platform. Simulation Platform is a cloud of virtual machines running GNU/Linux. On a virtual machine, various software including developer tools such as compilers and libraries, popular neural simulators, and scientific software such as Gnuplot, R and Octave, are pre-installed. When a user posts a request, a virtual machine is assigned to the user, and the simulation starts on that machine. The user can remotely access the virtual machine through a web browser and carries out the simulation interactively (a screenshot is shown in Fig. ​Fig.1).1). There is no need to install any software. It only requires a web browser. Therefore, Simulation Platform provides an ubiquitous computing environment for computational neuroscience research so as to free neuroscientists from tedious computer administration tasks and allow them to solely concentrate on their science. A demo site is open at http://sf4.sim.neuroinf.jp/~tyam/cns11/. Figure 1 A screenshot of a web browser during a computer simulation.

Effect of surrounding blur on foveal visibility.

Visibility of a simple stimulus is known to be determined not only by its physical contrast, but also by the configuration of surrounding stimuli. In this study, we investigated the surrounding modulation of foveal visibility of a blurred target. Subjects were instructed to respond to the gap orientation of a Gaussian-blurred Landolt ring presented at a fixation point with a surrounding stimulus. The correct response rate was measured as a metric of the foveal visibility. Results were subsequently compared among different surrounding stimulus conditions. Results showed an improvement in the subjects’ performance when low-pass white noise filtered with the same Gaussian function used for the target was presented in the surrounding area, although no effect was observed using high-contrast white noise. A performance improvement was observed when the surround stimulus had an intermediate contrast in the spatial frequency band necessary for identifying the target orientation.