The locations of active brain areas can be estimated from the magnetic fiel
d the neural current sources produce. In this work, we study a visualizatio
n method of magnetoencephalographic data that is based on minimum l(1)-norm
estimates. The method can represent several local or distributed sources a
nd does not need explicit a priori information. We evaluated the performanc
e of the method using simulation studies. In a situation resembling typical
magnetoencephalographic measurement, the mean estimated source strength ex
ceeded baseline level up to 2 cm from the simulated point-like source. The
method can also visualize several sources, activated simultaneously or in a
sequence, which we demonstrated by analyzing magnetic responses associated
with sensory stimulation and a picture naming task. (C) 1999 Academic Pres
s.