| CPC A61B 18/1492 (2013.01) [G16H 50/20 (2018.01); A61B 2018/00577 (2013.01); A61B 2018/00827 (2013.01); A61B 2018/00875 (2013.01); A61B 2018/1467 (2013.01)] | 20 Claims |
|
1. A computer-implemented method to find tissue proximity indications, performed when a catheter is inserted into a body part of a living subject such that electrodes of the catheter contact tissue at respective locations within the body part, the method comprising steps performed by a processor comprising:
receiving signals detected by the electrodes;
selectively rewarding and penalizing a reinforcement learning agent over reinforcement learning exploration phases to learn at least one tissue proximity policy responsively to at least one of the received detected signals;
applying the reinforcement learning agent in reinforcement learning exploitation phases to find respective tissue-proximity actions to be taken that maximize respective expected rewards responsively to the at least one tissue proximity policy, the applying includes applying the reinforcement learning agent in respective ones of the reinforcement learning exploitation phases to find respective ones of the tissue-proximity actions to be taken that maximize respective expected rewards responsively to respective states of the reinforcement learning agent and respective sets of available tissue-proximity actions to be taken;
providing respective derived tissue-proximity indications of proximity of a given one of the electrodes with the tissue responsively to the found respective tissue-proximity actions,
wherein the selectively rewarding and penalizing includes selectively rewarding and penalizing the reinforcement learning agent over the reinforcement learning exploration phases responsively to data of respective last ones of the reinforcement learning exploitation phases, the data comprising: respective ones of the states, respective found ones of the tissue-proximity actions to be taken, and respective actual rewards;
the method further comprising:
computing the respective impedance value of the given electrode for each one of the respective states responsively to at least one of the received detected signals;
computing the respective reference tissue-proximity indications independently of applying the reinforcement learning agent responsively to at least one of the received detected signals;
configuring a first set of the electrodes of the catheter to be electrically coupled to a first signal processing unit, wherein the computing the respective reference tissue-proximity indications is performed by the first signal processing unit responsively to receiving at least one signal provided by the given electrode;
configuring a second set of the electrodes of the catheter and the given electrode to be electrically coupled to a second signal processing unit, the second set of the electrodes being different from the first set of the electrodes, the given electrode being in the first set of the electrodes, wherein the computing the respective impedance value of the given electrode for each one of the respective states is performed in the second signal processing unit responsively to receiving the at least one signal provided by the given electrode, the first signal processing unit and the second signal processing unit providing input connections for the electrodes of the catheter, wherein the given electrode is electrically coupled to both the first signal processing unit and the second signal processing unit, the first signal processing unit processing tissue proximity indicators in a different way to the second signal processing unit;
computing respective impedance values of respective ones of the second set of electrodes by the second signal processing unit;
applying the reinforcement learning agent to find respective tissue-proximity actions to be taken that maximize respective expected rewards for the respective ones of the second set of electrodes responsively to the computed respective impedance values of the respective ones of the second set of electrodes; and
providing respective derived tissue-proximity indications of proximity of the respective ones of the second set of electrodes with the tissue responsively to the found respective tissue-proximity actions to be taken for the respective ones of the second set of electrodes,
the processor providing the respective derived tissue-proximity indications of proximity of the respective ones of the second set of electrodes connected to the second signal processing unit, that reconcile with results provided by the first signal processing unit.
|