This chapter introduces a newly developed integrative fMRI-MEG method combined with a spatial filtering (beamforming) technique as a non-invasive neuroimaging method to reveal dynamic processes in the brain. One difficulty encountered when integrating fMRI-MEG analyses is mismatches between the activated regions detected by fMRI and MEG. These mismatches may decrease the estimation accuracy, especially when there are strong temporal correlations among activity in fMRI-invisible and -visible regions. To overcome this difficulty, a spatial filter was devised based on a generalized least squares (GLS) estimation method. The filter can achieve accurate reconstruction of MEG source activity even when a priori information obtained by fMRI is insufficient. In addition, this chapter describes the feasibility of a newly developed optically pumped atomic magnetometer as a magnetic sensor to simultaneously measure MEG and MR signals.
TopIntroduction
What is the mind? What mechanisms in the brain are associated with visual awareness? An important step toward answering these questions is obtaining precise knowledge about the dynamic brain processes involved in these functions. Although recent neuroimaging techniques such as magnetoencephalography (MEG) (Hämäläinen, Hari, Ilmoniemi, Knuutila, & Lounasmaa, 1993), positron emission tomography (PET), near-infrared spectroscopy (NIRS), and functional magnetic resonance imaging (fMRI) (Frackwiak, Berkinblit, Fookson, & Poizner, 1998; Murata & Iwase, 2001; Augustyn, & Rosenbaum, 2005; Adam, Mol, Pratt, & Fischer, 2006; Kovacs, Buchanan, & Shea, 2008) have become powerful tools for exploring higher brain functions (Kobayashi, Ozaki, Nagata, 2009), each technique has limited spatial and/or temporal resolution that hamper our understanding of dynamic brain processes.
To overcome these limitations, neuroimaging methods that fuse multimodal techniques are being developed (Dale, Liu, Fischi, Buckner, Belliveau, Lewine, & Halgren, 2000; Schulz, Chau, Graham, McIntosh, Ross, Ishii, & Pantev, 2004; Okamoto, Dan, Shimizu, Takeo, Amita, Oda, Konishi, Sakamoto, Isobe, Suzuki, Kohyama, & Dan, 2004; Carrie, Reynolds, Goodyear, Ponton, Dort, & Eggermont, 2004). However, at present, there is no applicable technique that can provide sufficiently high spatial and/or temporal resolution. We have developed an integrative fMRI-MEG neuroimaging method to analyze the dynamic activation of multiple cortical areas (Innami, Kobayashi, Jung, Ohashi, Hamada, Nagamine, Fukuyama, Azuma, & Tsutsumi, 2004; Ohashi, Innami, Jung, Hamada, & Kobayashi, 2006; Okada, Ohashi, Jung, Hamada, & Kobayashi, 2007). Here, we introduce the latest version of the fMRI-MEG integrative neuroimaging method.
MEG (with superconducting quantum interference devices, SQUIDs) and high-field MRI (with superconducting magnets that require cryogenic cooling) are difficult to measure simultaneously. Optically pumped atomic magnetometers (OPAMs) are currently expected to overtake SQUIDs, and the possibilities for using OPAMs for biomagnetic field measurements and MRI have been demonstrated. We have developed a highly sensitive atomic magnetometer as a magnetic sensor to measure both MEG and MR signals. We describe the principles of the atomic magnetometer and the results of biomagnetic field measurements.