RECURSIVELY-APPLIED SCANNING ALGORITHMS FOR INVERSE ANALYSES OF GASTROINTESTINAL BIOMAGNETIC FIELDS

摘要

The process of identifying dipolar sources of bioelectricrncurrent in the gastrointestinal (GI) tract is of greatrnimportance to the future of medical diagnosis.rnSuperconducting QUantum Interference Device (SQUID)rnmagnetometers measure the minute magnetic fieldsrnassociated with electrical control activity (ECA) from GIrnsmooth muscle. The phenomenology of abnormal ECArncurrent propagation can offer important insights into thernearly identification and treatment of many diseases,rnincluding gastroparesis and ischemia. In order to identifyrnthese gastrointestinal sources of current, a biomagneticrnfield reconstruction program has been developed in whichrnthe abdomen is modeled using a set of dipoles that arerndistributed so as to produce an anatomically realisticrnsimulation. A numerical algorithm for solving the inversernbiomagnetic problem was developed, in which a novelrndipole localization procedure was implemented in order tornidentify dipole sources of current based on magnetic fieldrnvalues recorded using SQUID magnetometers. Thisrncomputational inverse reconstruction was tested withrnexcellent results on simulated data sets containing 1, 3,rnand 190 dipoles. In all cases, dipoles were identified atrncorrect locations, consistent with propagating GIrnelectrical activity, and with dipole moments that oscillatedrnwith frequencies accurately corresponding to the knownrnECA frequencies. These results are very promising forrnfuture studies of abnormal gastrointestinal propagationrnphenomena, and for the development of noninvasivernmedical diagnosis procedures for the stomach andrnintestine.