Magnetoencephalographic data indicates separate component modeling is necessary in evaluating evoked responses

Abstract

A double Gaussian Model was used t o estimate dipole source contributions t o magnetoencephalograms. The method w a s tested using four 2mn length dipoles fixed in conductive material enclosed within the human s k u l l . simulated by Gaussian activation of one dipole with a mean latency of 100ms, and a second dipole a t 200ms. temporal bone and the occipital bone. Source contributions t o each resultant waveform were estimated as the amplitude parameters of a best f i t double Gaussian function. Canparing dipole inverse solutions frcm waveform paks with those frun estimated contributions revealed only a mall difference in localization accuracy. Source strength is s l i g h t l y better estimated using the component d e l method. The peak picking method overestimated strength when dipoles were aligned and underestimated when they were opposed. hroked responses were Simulations were performed over the l e f t 1NI"CTION cXlr laboratory measures auditory and visual A evoked responses t o study the behavior of cognitive components under various workload conditions. typical evoked f i e l d w i l l have several canponents occurring a f t e r stimulus onset. of various canpnents typically localize the neural source contributing t o the component a t a unique location (1,2,3). The localization of a 100 ms auditory component has a different location f r m that of a 160 ms (31, and thus can not be modeled simply as a generator reversing polarity. In t h i s paper, a Gaussian model is used t o simulate many closely packed neurons being activated with distributed latencies. 'Ihe model might be used t o estimate neural sources' contributions t o the canponents of the observed evoked response. Inverse solutions Four current dipoles were constructed frun 32 Pm; single conductor shielded wire. insulation and the shield were cut back t o give a 2nnn length between the exposed center conductor and The outer the shield. the ends of the dipole. conductor simulates intracellular currents of many closely packed c o r t i c a l neurons. The dipoles were positioned i n a s o f t sponge, wrapped in chamois cloth, and soaked i n a conductive electrode paste solution. material allowed the return currents t o better approximate extracellular volume currents. !&a dipoles were placed i n the l e f t tempxal region t o simulate auditory evoked responses (AEX) and the other two were placed i n the occipital region t o simulate visual evoked responses (VER).After enclosing the simulated brain within the s k u l l , x-ray images were used t o precisely locate the true dipole positions. An x-ray shot was taken parallel t o each axis of a designated "head reference" coordinate system. The x-axis passes through s k u l l l a d m a r k s j u s t above the external ear or i f ices, the y-axis passes between the nasd bones, and the z-axis is the cross product of x and y. Because the x-ray procedure w i l l produce an expanded image as a function of cbject position between the x-ray source and t h e film, a ccmputer program was written t o convert each...