P 218. Optimized small animal transcranial magnetic stimulators using distributed coils

Abstract

Question The precise impact of TMS on the neural pathways and the mechanisms of action remain unknown. Performing studies of different coil configurations in human beings is restricted due to ethical reasons and it is difficult to gather statistically significant data of large human study groups. Therefore, to explore TMS in a systematic and flexible way, miniaturization of TMS for rodent brain studies is a complementary addition to the human studies. Current stimulators for small animal studies lack a high degree of focus of electric field. More optimal designs are thus needed. Methods We aim at numerically optimizing a distributed planar coil array that can generate a focal electric field in rodent brains. The distributed coil array consists of N × N coils distributed in a rectangular grid (34 mm × 34 mm). Since it is practically difficult to activate each coil with a different current, we aim at the activation of the coil array with a single current by placing the different coils in series connection. Since only a single current is flowing through the coil array, the geometrical configuration, e.g. radii and number of turns, of the different coils needs to be optimized. The numerical optimization is performed using a genetic algorithm ( Crevecoeur et al., 2010 ) on the basis of the calculated electric fields. The electric fields are calculated starting from the magnetic fields. Further details on the electrical field solver can be found in ( De Geeter et al., 2012 ). The rat head is modeled as four concentric ellipsoids representing the tissues scalp, bone, cerebrospinal fluid (CSF) and brain. The minor and major axes of the inner layer brain are 14 mm and 28 mm long and the other layers are 1 mm and 2 mm thick near the minor and major axes respectively, similar to the rat’s real dimensions. Results We performed simulations onto a rectangular coil array with N = 2 and N = 4. We observed that N = 2 yields a limited amount of degrees of freedom so to excite a focal volume of high electric field. N > 4 is practically very difficult to implement because the radii of the coils would become too small. The optimal spatial electric field distribution in the sagittal plane is given in the figure here below for the N = 4 distributed coil array (right). As a comparison, a non-optimized electric field distribution (left) is also given, illustrating the impact of the coil configuration parameters. In this configuration we fixed the outer radii of the coils to be 4 mm, so that the coils can be placed within the rectangular grid, as well as the current direction in the different coils. The current directions were chosen similar to the current directions in the figure-8 coils. The optimization parameters were here the number of turns. About 500 genetic algorithm generations were needed for the optimization which resulted in approximately 8 h computation time. Conclusions An optimized small animal TMS is presented that enables focal stimulation by activating localized high electric fields. Simulations were performed on a simplified rat model and can be extended towards more realistic head models of rats. Moreover, the degrees of freedom in the design can be extended by including unknown radii of the coils. The numerically designed TMS can be deployed for further in-depth in vivo studies that evaluate the impact of stimulation using neuro-imaging techniques.