Muneo ONO,¹ Yoshihiko KASHIO,¹ Martin SCHWEIGER,² Hamid DEHGHANI,³ Simon R. ARRIDGE,² Michael FIRBANK⁴ and Eiji OKADA¹

¹Department of Electronics and Electrical Engineering, Keio University, 3-14-1, Hiyoshi, Kohoku-ku, Yokohama, 223-8522 Japan, ²Department of Computer Science, University College London, Gower Street, London, WC1E 6BT, UK, ³Department of Medical Physics and Bioengineering, University College London, Capper Street, London, WC1E 6JA, UK, ⁴University Department of Radiology, Imaging Laboratory, Royal Victoria Infirmary, Newcastle upon Tyne, NE1 4LP, UK

(Received May 25, 2000; Accepted June 30, 2000)

An accurate modelling of light propagation in the head is required to develop an algorithm to reconstruct the image of brain activity. Most previous studies have calculated the light propagation in two dimensional models because of their advantage in computation time and memory requirement over three dimensional models. However, in topographic imaging, the sensitivity distribution in the cross sections parallel to the brain surface which cannot be obtained from a two dimensional model is most important to reconstruct the image. In this study, the light propagation in three dimensional adult head models is calculated by finite element method and hybrid radiosity-diffusion method. The light propagation in the adult head is strongly affected by the non-scattering cerebrospinal fluid (CSF) layer surrounding the brain. The sensitive area is shifted toward the deeper region, and is spread around the CSF layer. The intensely sensitive region on the brain surface is broadly distributed between the source and detector. However, the sensitive region does not penetrate into the deeper part of the brain.

Key words : near infrared spectroscopy, diffusion equation, finite element method, hybrid radiosity-diffusion model, topographic imaging, photon measurement density function

okada@elec.keio.ac.jp