Programmer and associated method for use in sensing vector selection with an implantable medical device

p00001Programmer and associated method for use in the selection of detection vectors with an implantable medical device.

p00002Technical Field The present invention relates to the technical field of implantable medical devices. More specifically, the present invention relates to implantable medical devices that detect cardiac activity.

p00003Background Implantable cardiac stimulation devices often include electrical cardiac activity detection circuits to observe a patient's cardiac function.

p00004Document US2005 / 049644A1 refers to an implantable cardiac treatment system that is provided with electrodes arranged in various positions in a patient's chest. During system operation, various detection vectors can be monitored periodically, repeatedly or continuously to select a detection vector used for the analysis.

p00005Document W099048554A refers to a system and a medical device method for plotting symbols that represent, on an interactive display, complexes of selected arrhythmia episodes. By visualizing on an interactive display screen the episodes of cardiac arrhythmia recorded by a patient, the doctor can evaluate and interpret the nature of the cardiac arrhythmias of the patient and provide a schedule of the patient's ICD.

p00006WO / 2004/105871 refers to a detection architecture for use in implantable heart rate devices. The respective detection architecture provides methods and devices to discriminate between arrhythmias. In addition, taking advantage of the improved specificity at the origin of the arrhythmia identified, the detection architecture can discriminate between appropriate rhythms for treatment of the device and those that are not.

p00007WO / 2004/108212 refers to a detection architecture for use in heart rate management devices. The detection architecture provides a method and means to certify the episodes detected by the heart rate management device. In addition, taking advantage of the enhanced ability to accurately identify only the detected episodes that are desirable, and preventing the use of episodes marked as suspicious, the detection architecture can discriminate between appropriate rhythms for treatment of the device and those that are not.

p00008New and alternative methods and devices that are adapted to identify and / or select favorable detection vectors are desired.

p00009Summary The present invention is defined by the features of the claims.

p00010The present invention, in illustrative examples, includes methods and devices configured to analyze detection vectors in an implantable cardiac stimulation system. In an illustrative example, a first detection vector is analyzed to determine if it is suitable, within given threshold conditions, for use in detecting and analyzing cardiac events. If so, the first detection vector can be selected for detection and analysis. Otherwise, one or more additional detection vectors are analyzed. In another illustrative embodiment, a set of detection vectors are analyzed instead of possibly identifying a suitable vector without analyzing additional vectors. A polynomial formula or other mathematical formula can be used during the analysis to generate a metric that indicates the suitability of the detection vector for use in detecting and analyzing cardiac events. In another example, a query table or diagram can be used to generate a metric that indicates the suitability of the detection vector. A detailed illustrative example includes methods for analyzing detection vectors using an index assignment system. Additional illustrative examples include systems and devices adapted to perform at least these methods. One such illustrative example includes an implantable medical device, which can be an implantable cardioverter / defibrillator, adapted to perform these methods. Another example includes a programmer configured to perform these methods, including certain steps of directing the operation of an implanted or implantable associated medical device. Another additional illustrative example includes a system that has an implantable device and a programmer, in which each of the devices collaboratively performs the steps of carrying out the analysis of the detection vectors.

p00011Brief description of the drawings

p00012Figures 1A and 1B, respectively, show cardiac stimulus systems implanted so

p00013subcutaneous and transvenous in relation to the heart;

p00014Figure 2 is a block diagram illustrating an initialization method for an implantable system of heart stimulation; Figures 3A and 3B are graphical representations of cardiac signals, illustrating an analytical form for identify QRS and T waves; Figure 4 is a graph showing the treatment of a cardiac signal, for explanatory purposes; Figures 5A and 5B illustrate a block diagram for a method of vector signal analysis; Figure 6 illustrates a simplified model of an illustrative method; Figures 7A and 7B illustrate a block diagram for a signal analysis method, within the analysis vector of figures 5A and 5B; Figure 8 illustrates a block diagram for a method of signal quality assessment cardiac; Figures 9A to 9C illustrate a block diagram for a method of analyzing a cardiac signal; Figures 10A to 10C are graphs illustrating mathematical relationships used in illustrative embodiments; Figures 11A and 11B illustrate a block diagram for a method of vector signal analysis; Figure 12 is a block diagram illustrating a method in which a detection vector is selected. primary and a secondary detection vector for use in detecting and analyzing cardiac episodes; Figure 13 is a block diagram for an illustrative embodiment; and Figure 14 is a block diagram for another illustrative embodiment.

p00015Detailed description The following detailed description should be read with reference to the drawings. The drawings, which are not necessarily to scale, represent illustrative embodiments and are not intended to limit the scope of the invention.

p00016Figures 1A and 1B, respectively, show cardiac stimulus systems implanted subcutaneously and transvenously in relation to the heart. Referring to Figure 1A, the heart 10 of a patient is shown in relation to an implanted, subcutaneous cardiac stimulation system that includes a container (rec) 12. A conductor 14 is secured to the container and includes a detection electrode A 16, a coil electrode 18 and a detection electrode B 20. A rec electrode 22 is shown on the container 12. Therefore, several detection vectors are available, including A-rec, B-rec and AB. It should be understood that each pair of electrodes actually forms a first vector of a first polarity and a second vector of a second polarity, opposite; for convenience, reference is simply made to the combination of electrodes. It should be noted that it is also possible to use coil electrode 18 as a detection electrode. In addition to detecting cardiac activity through cardiac electrical signals, some or all of the electrodes can be used to detect respiration or other activity or physiological state. Illustrative subcutaneous systems are shown in US Patent Nos. 6,647,292 and 6,721,597. Some embodiments include a unitary system that has two or more electrodes on a housing as set forth in patent 6,647,292, instead of the one shown in Figure 1A. A unit system that includes an additional conductor can also be used.

p00017Referring now to Figure 1B, a transvenous system is shown relative to the heart 30 of a patient. The transvenous cardiac stimulus system includes a container 32 connected to a conductor 34. The conductor 34 enters the patient's heart and includes electrodes A 36 and B 38. Additional electrodes may also be included for stimulus detection or delivery, and can be also be used for detection in some embodiments of the present invention. In the illustrative example, electrode A 36 is generally located in the patient's ventricle, and electrode B 38 is generally located in the patient's atrium. The conductor 34 may be anchored in the patient's myocardium. The conductor 34 may also include one or more coil electrodes, internal or external to the heart, as shown with the number 39, which can be used to provide stimulation and / or to detect cardiac or other activity, such as breathing . A rec electrode 40 is shown on the container 32. With this system, multiple detection vectors, of first and second polarities, can also be defined. In both figures 1A and 1B, one or more detection electrodes can also be used for stimulus delivery. Some embodiments of the present invention may be used in combination systems that may include defined detection vectors between two subcutaneous electrodes, a subcutaneous electrode and a transvenous electrode, or two transvenous electrodes.

p00018In the configurations of Figures 1A and 1B, multiple detection vectors are available. The detection of cardiac function together with at least one of these detection vectors allows the implanted cardiac stimulation system to determine if the treatment is indicated due to the detection and identification of an unfavorable prognosis condition such as, for example, a tachycardia ventricular An implant doctor can make a selection of vectors by directly diagnosing which of the captured vectors is the best. However, this requires an evaluation of cardiac function along with several vectors and may increase the time needed to perform the implantation, and also increases the risk of human error. In addition, if the doctor needs to make a selection of vectors, as the patient's physiology changes (which can happen while healing develops around an implanted detection electrode), the system runs the risk of using a detection vector. suboptimal until the patient visits the doctor again. Finally, the selection of a vector is often a task that requires specialized training, since the selection of a suitable vector from those available is not necessarily intuitive.

p00019A robust method of selecting detection vectors, as well as devices adapted to perform such methods, is desirable. The present invention, in illustrative embodiments, provides such methods and uses various criteria to do so. Some embodiments include implantable devices and programmers for implantable devices that are adapted to perform such methods.

p00020The systems shown in Figures 1A and 1B may include operating circuits and a power source housed within the respective containers. The power source can be, for example, a battery or a battery bank. The operating circuits may be configured to include said controllers, microcontrollers, logical devices, memories, and the like, as selected, needed or desired to perform the illustrative methods set forth herein. The operating circuits may also include (but not necessarily) a load subcircuit and an energy storage subcircuit (for example, a capacitor bank) to generate a stored voltage for cardiac stimulus that is in the form of a cardioversion stimulus and / or defibrillation The operating circuits may also be adapted to provide a cardiac electrostimulation output. Both the cardioversion / defibrillation and cardiac electrostimulation subcircuits as well as the capabilities may be incorporated into a single device. The methods described below may be contained in a hardware within the operating circuits and / or as instruction sets for operating the operating circuits and / or in the form of machine-readable media (optical, electrical, magnetic, etc.) containing such instructions and instruction sets.

p00021Each of the devices 12, 32 may further include such components, as would be appropriate for communication (such as RF communication or inductive telemetry) with an external device such as a programmer. For this purpose, programmers 24 (figure 1A) and 42 (figure 1B) are also shown. For example, during an implantation procedure, once the implantable device 12, 32 and the conductors (if included) are placed, the programmer 24, 42 can be used to activate and / or direct and / or observe diagnostic tests or operational. After implantation, the programmer 24, 42 can be used to non-invasively determine the status and background of the implanted device. The programmer 24, 42 and the implanted device 12, 32 are adapted for wireless communication, allowing the implanted device to be asked. The programmers 24, 42 in combination with the implanted devices 12, 32 may also allow the indication of statistical data, errors, background and potential problems for the user / doctor.

p00022In some embodiments, implanted devices 12, 32 directly perform the following illustrative methods independently (periodically or occasionally) or in the direction of a programmer 24, 42. In other embodiments, the following methods are performed using implanted devices 12, 32 to carry out data collection, the programmer 24, 42 performing other analytical steps by downloading captured data (for example, in real time, or in blocks of default size). The programmer 24, 42 can press the appropriate data collection for the following illustrative methods, and the programmer 24, 42 can then direct the implanted device 12, 32 to use a selected vector. Some methods can be performed by distributing appropriate tasks between the implanted device 12, 32 and the programmer 24, 42.

p00023Figure 2 is a block diagram illustrating a method for detecting the initialization of an implantable cardiac stimulation system. For the vectorization method, a detection threshold for the implantable cardiac stimulation system is initialized by performing a detection process with one or more available detection vectors. For example, the steps of the method shown in Figure 2 can be performed to detect the vectors A-Rec, B-Rec and AB, shown in Figures 1A and 1B. The steps can also be performed using vectors that include the shock / coil electrode 18 shown in Figure 1A, as well as any stimulus electrode in Figure 1B, although the assembly under consideration is reduced for illustrative purposes in the following examples to A -Rec, B-Rec and AB.

p00024A vectorization method is a method to analyze one or more detection vectors available in order to select a detection vector for use in the detection of cardiac events. The method in Figure 2 is used to establish the detection parameters for a vectorization method. Figure 2 includes first setting an initial lower limit of detection, as shown with the number 50, preferably in a relatively low range (high sensitivity) so that the device will detect cardiac episodes that exceed the lower limit of detection. For example, said initial lower limit of detection can be set at a percentage of a historical value measured by several previously detected cardiac episodes, specific to the patient, a patient population, a particular implantation configuration or other suitable variables.

p00025As shown in block 52, several iterations of a sub-method are performed. In block 52, an episode is detected, as shown with the number 54, and then the lower detection limit is raised, as shown with the number 56. The episode detection 54 may occur when the detected signal crosses the lower limit of detection, and may be followed by a refractory period during which detection is rendered useless. The iterations in block 52 may continue until a disconnection time is presented, as shown with the number 58. The disconnection time 58 may occur, for example, when a period of 2 seconds ends without an occurrence of detection. The disconnection time may indicate that the lower limit of detection has been raised above the intensity of the signals received for cardiac events.

p00026After the disconnection time at 58, the lower limit of vectorization detection is set to a percentage of the lower detection limit that led to the disconnection time 58, as shown by the number

p0002760 For example, the lower limit of vectorization detection can be set at approximately 40% of the level that led to the disconnection time. Next, time periods are set, as shown with the number 62, and the vectorization analysis is performed as shown with the number 64. The time periods are explained with reference to Figures 3A and 3B. The lower limit of vectorization detection can be set differently for each of the detection vectors being analyzed, since each vector can produce a different signal intensity from the other vectors. For example, the method shown in Figure 2 can be repeated for each of several vectors before performing any vectorization analysis 64, or the method of Figure 2 can be performed as a part of the vectorization analysis for each individual vector.

p00028Figure 3A is a graphical representation of a heart signal illustrating an analytical way to identify the QRS and T waves. A heart signal is shown with the number 80. A vectorization detection threshold is shown with the number 82. A episode at 84, when cardiac signal 80 crosses the threshold of vectorization detection 82. A refractory period 86 is presented following detection 84. After the refractory period 86, one or more extension time periods (CT) are presented, as shown by the number

p0002988. The peak that occurs during the refractory period 86 is assumed to be, at least initially, the R wave, as shown by the number 90. A peak value of the signal is identified during the CT period, as shown by the number 92, and it's supposed to be noise. This peak 92 may be the T wave, but it is not necessarily so. In an illustrative example, the refractory period 86 lasts 160 milliseconds, and the period or periods CT 88 last 220 milliseconds, although these values ​​may vary in other embodiments.

p00030In the illustrative example, it is assumed (for vectorization purposes) that the longest QT interval will be approximately 400 milliseconds and, thus, the peak value detected during the period CT 88 is assumed to be the T wave. The illustrative method, with the parameters given above, is adapted for use with a patient who has a heart rate of 30 to 150 beats per minute (bpm). The detection threshold 82 is effective during the CT 88 period, so that if the detected signal crosses the detection threshold 82 during the CT 88 period, this will be treated as another detection.

p00031Figure 3B illustrates that a new detection 94 will be defined for a threshold that crosses during the period CT 88A following a first refractory period 86A associated with a first detection 96. The new detection 94 has its own refractory period 86B and its period CT 88B. Given a frequency of 150 bpm or less, the detections shown in Figure 3B indicate a double detection, since the detections are too close to each other. It can be seen that, when double detection occurs, detection outside of periods CT 88A, 88B may not occur, although this is not always the case. As explained below, the present methods are adapted to select at most one of the detections 94, 96 as representative of a heartbeat, while determining that the other detection 94, 96 is noise.

p00032Since different detection vectors can detect episodes in different ways, and since the target population for such devices often has abnormal cardiac function, a prototypical PQRST form may not always be present, as shown in Figure 3A For example, the T wave may be relatively larger than the R wave. Some of the present methods also include an ability to establish an indicator that requests that an attending physician (the "operator") observe the heart signal and determine whether the R wave or the T wave has a larger amplitude. When certain conditions are met, the indicator is set and the operator can be asked to enter it.

p00033Figure 4 is a graph showing the treatment of a heart signal for explanatory purposes. During the following vectorization example, the captured heart signal is marked to identify episodes. The signal, shown with the number 100, is detected in relation to a detection threshold 102. When the signal 100 exceeds the detection threshold 102, this is marked as an episode. The most recent chronologically episode is marked as episode i, as shown with the number 106, while an immediately preceding episode is marked i-1, as shown with the number 104. The duration between detections 104 and 106 is defined as interval , while the amplitude of peak 106 is the picoi, and the height of the noise peak following peak 106 is noisei. One objective is to define the episodes and sections detected in the cardiac signal in the manner shown in Figure 4; The methods of Figures 5A and 5B, as well as other methods shown below, are adapted to achieve this objective, when possible.

p00034Figures 5A and 5B illustrate a block diagram for a method of detection vector analysis. The illustrative method begins at 130. A first vector is analyzed, as shown with the number 132. The details of the analysis are further explained elsewhere. This analysis may include performing the method of Figure 2, or the thresholds defined in Figure 2 may be defined before performing the method for the first vector or more vectors.

p00035The analysis at 132 may provide an Index, as shown with the number 136, or it may not provide an Index, and instead provide Possible Indices, as shown with the number 134.

p00036The use of the terms "Index" and "Possible Indices" should be explained. In an illustrative example, the vectorization analysis looks at parameters of the detected heart signal to determine if the signal is likely to be useful for the detection of cardiac events. In an illustrative example, the analysis of intervals and / or amplitudes is used to determine if "good" detections are occurring. A "good" detection can have a desirable signal-to-noise ratio (SNR) and / or amplitude range, and can prevent the detection of "unwanted" episodes "(variations). If the analysis indicates that regular detections are occurring, and that premature detections are not occurring, an Index for the detection vector is calculated. The Index is an example of a metric for the received signal that indicates the quality of the detection vector for the purpose of detecting cardiac events. In an illustrative embodiment, the Index is calculated using a polynomial formula that includes terms that refer to SNR and / or the amplitude of the signal or signals captured. In another illustrative embodiment, the Index is calculated using a query table or diagram that can again refer to the SNR and / or the amplitude of the signal or signals captured.

p00037If the analysis does not analyze, with certainty, noisy detections regarding detections of cardiac events, then, two or more Possible Indices can be calculated, each based on an assumed resolution of the ambiguity or uncertainty that prevents the calculation of an index. For example, two Possible Indices can be calculated, one assuming that the QRS peak exceeds the T wave or another noise peak, and the other assuming that the QRS peak does not exceed the T wave or other noise peak. The following illustrative examples use a particular calculation of Indexes and Possible Indexes to detect vectors, the details of which can be modified in various suitable ways. It is sufficient that the Index or other metric provide an indication of the quality of the signal captured together with a detection vector for the purpose of detection and / or analysis of cardiac events.

p00038Returning to the example in Figure 5A, if the analysis only provides Possible Indices, as indicated by the number 134, an interrogation indicator is set for the first vector, as shown with the number 138. The interrogation indicator indicates that the The operator may have to provide an input to the devectorization process to determine which of the Possible Indices is the correct Index to use with the first vector observing whether "QRS> T wave?". However, instead of immediately asking the operator for help, the method instead analyzes the second vector, as shown with the number 140.

p00039Returning to step 136, if an Index is provided for the first vector, the method then determines whether the Index for the first vector exceeds a predetermined threshold, as shown with the number 142. In the illustrative method, if the threshold is exceeded , this indicates that the first vector provides excellent detection, and any additional analysis is considered unnecessary. Therefore, if the threshold for the first vector is exceeded, the method ends by selecting the first vector, as indicated by the number 144. Otherwise, the method continues to analyze the second vector, as shown by the number 140. .

p00040Steps 142 and 144 represent a premature exit condition for the system, which may allow it to terminate the selection of vectors before a "good" vector is identified. In another embodiment, as described below with reference to Figure 14, premature exit conditions can be omitted and / or disabled. The system would then be configured to test each vector before making a selection of one vector with respect to the others. In an illustrative example, a vector with the highest index can be chosen when premature exit conditions are disabled or unavailable. In another illustrative example, each vector is analyzed, and the selection of vectors is presented in a preferred manner, in which a first "preferred" vector is compared with a threshold similar to that of steps 142 and 144 after analysis. of each vector.

p00041As in the case of the first vector, the analysis of the second vector provides an Index, as indicated by the number 152, or a Non-Possible Index and, instead, provides Possible Indices, as shown by the number

p00042150 If Possible Indices are provided, the interrogation indicator is set for the second vector, as shown with the number 154. The method then goes to A, as shown with the number 156, which continues in Figure 5B.

p00043If the analysis of the second vector provides an Index, the method follows from block 152 to determine if the index for the second vector exceeds a threshold, as shown with the number 158. Again, if the threshold is exceeded, this indicates that the second vector provides excellent detection and, thus, the method ends by selecting the second vector, as shown with the number 160. As before, the premature exit conditions of steps 158 and 160 can be disabled or omitted. Otherwise, the method continues to A 156 in Figure 5B. The steps shown for the first and second vectors can be repeated for any number of vectors, depending on the particularities of the implantable system. heart stimulation

p00044Referring to Figure 5B, from A 156, the method determines whether any of the interrogation indicators has been established, as shown with the number 162. If not, the method simply selects the detection vector with the best Index, as shown with the number 164. If one or more indicators have been established, the method continues at 168 analyzing the Possible Indices. Questions may be asked of the operator (attached physician) in relation to any of the Indices or the Possible Relevant Indices.

p00045The analysis in 168 may include determining if any questions have to be asked. In one embodiment, if any vector provides an Index, no question is asked of the operator. In this way, the operator simply allows the implanted device to select a detection vector unless the input is, in fact, necessary.

p00046In another embodiment, if at least one vector provides an Index, and none of the Possible Indices available from other vectors exceeds all available indices, the operator is not asked any questions, since the Possible Indices cannot provide the best available vector. Only relevant questions can be asked by selecting the vector that has the best possible Index first. If needed, the operator is asked the question or questions, as shown with the number 170. If the operator's response eliminates the best Possible Index, the method can be iterated to ask additional questions regarding one or more of the best Possible Indexes following. Or, if the answer eliminates the best possible index, and the highest remaining index is an index, instead of a possible index, a corresponding detection vector is selected. Next, the detection vector with the best Index is selected, as shown with the number 164.

p00047In another additional embodiment, the implanted device can initiate the vector selection process by itself, when it is not in communication with the programmer. This may be the case, for example, if the device determines, during operation, that it cannot accurately observe cardiac activity using a selected vector. If the implanted device starts the selection of vectors when it is not in contact with a programmer, it cannot request or receive an input from the user. In this example, if no vector index can be generated, the device can re-observe the previously used vector, even if it has been found to be inadequate. This circumstance of malfunction can be treated by some embodiments in which the vectorization method identifies a first vector, primary or by default, as well as a second vector, reinforcement, as described further below.

p00048It should be noted, with respect to Figures 5A and 5B, that the part of the method shown in Figure 5 selects a vector depending on whether the vector provides a sufficient Index. The part of the method in Figure 5B, on the other hand, selects the vector with the best Index.

p00049In some embodiments, both first and second detection vectors are selected, the first detection vector being a primary or default detection vector and the second detection vector being an alternative or clarifying vector. For example, during a given cardiac episode, if the first vector provides, at best, ambiguous indications of whether treatment is needed, the second vector can be used to resolve any ambiguity. In other embodiments, a second detection vector can be identified during the initialization procedure in the event that, at a later time, the first detection vector becomes inadequate (due to changes in physiology, external noise, etc. .) or not available (due to mechanical failure, for example). For embodiments that identify the first and second detection vectors, the method of Figures 5A and 5B can simply follow, using similar steps and processes, until a second vector is similarly identified. A third vector and additional vectors can also be identified.

p00050An illustrative method is shown in Figure 6. A detected signal is analyzed using detection techniques in various significant characteristics including, for example, peak amplitudes, intervals between peaks, and noise levels between peaks. Figures 5A and 5B illustrate an analytical form for this first stage. Next, the characteristics identified by the detection techniques are further analyzed to reduce the number of variables under consideration. In the illustrative example of Figure 6, a set of {n} episodes detected is analyzed to generate a set of {a} identified QRS peaks and associated noise levels, in which the only known relationship between nya is that it may not be analyzed and condense

p00051ai Since the signal easily detected within the data pairs shown, ambiguity is admitted during signal processing also carrying more "additional data". Additional data may include temporary data indicating the time at which the detected episodes occur. If necessary, the implications of the “additional data” can be determined by looking for an entry by the user, as discussed further below.

p00052Figures 7A and 7B illustrate a block diagram for a particular form of the method of Figure 6. The example illustrated in Figures 7A and 7B is a particular way to transfer data related to a detection vector to metrics to evaluate the advantages of a given detection vector against one or more thresholds and / or other detection vectors. From the beginning block 200, the illustrative method identifies a detection threshold for vectorization, as shown with the number 202. In some embodiments, the method of Figure 2, for example, can achieve step 202. Next, the Illustrative method captures contiguous detected episodes, as indicated by the number 204. "Detected episodes" may occur when the detected signal crosses the detection threshold.

p00053Samples of selected data associated with the detected episodes are also saved for analysis. In an illustrative embodiment, n = 11, although other sets of larger or smaller episodes can also be used. An iterative analysis follows for episodes initially marked with the variable i, following the iterative method until i n, although in some embodiments the method may be suspended if it becomes apparent that the vector under consideration is inappropriate. For the first iteration, which comes from block 204, the illustrative method uses i = 1 and also sets another variable a = 1.

p00054The loop then begins for the {n} episodes detected considering the interval between a first episode (i-1) and a second episode (i). As shown with the number 206, an interval between episode i and episode i-1, Interval (i), is compared to a threshold, with 400 milliseconds used for illustrative purposes. If the threshold is not exceeded, the method continues at B, as shown with the number 210, in Figure 7B, unless, as indicated in step 208, the first interval was being considered (Interval (1) ), in which case the method is simply iterated to i = 2, as shown with the number 212, and again loop back to step 206. If, in 206, the interval is greater than the threshold, then the interval is stored as Interval (a), as shown with the number 214, and the peak captured during the refractory period i is stored as QRS (a), As shown with the number 216.

p00055The interval between ie i + 1 (Interval (i + 1)) is then considered as shown with the number 218, and compared with a threshold, which is shown again for illustrative purposes, such as 400 milliseconds. If the threshold is exceeded by 218, then the peak signal captured during CT (i), the extension time following the refractory for detection i, is stored as Noise (a), as indicated by the number 220. The method is then iterated as shown with the number 222, and returns to step 206 but with i = i + 1 already = a + 1.

p00056Returning to step 218, if the interval (i + 1) considered in that case does not exceed the threshold, then the refractory peak that caused this last detection is stored as Noise (a), as shown with the number 224. This it occurs because it is assumed that this last detection was presented too close to the previous detection to be another QRS complex. The illustrative 400 millisecond threshold interval is used in conjunction with a method that should be performed when the patient's heart rate is in the range of approximately 30 to 150 beats per minute (bpm). Detections that occur with less than an interval of 400 milliseconds would correspond to a larger heart rate (150 bpm or more), and the method therefore assumes that shorter intervals indicate that one or more detections are caused by noise. The method can be adapted to other patient's heart rates, if desired.

p00057Since 224, the next iteration begins at 226 after another iteration, this time with i = i + 2 ya = a + 1. The variable i receives a double iteration since, as shown in 224, the peak (i ​​+ 1) is considered to be noise. When returning to step 206, in some embodiments the following interval is taken from i-2 to i, which covers episode i-1 that has been identified as likely to occur due to noise. For example, at a microcontroller level, an indicator can be set to indicate a double iteration in step 226, the indicator being reset once the interval in step 206 is considered.

p00058Turning now to Figure 7B, the method chooses in block B 210, which follows from block B 210 in Figure 7A. As shown with the number 232, the method continues by determining whether a “long QT” condition has been verified. In an illustrative embodiment, the "long QT" condition allows an operator to indicate to the implantable device system and / or the programmer that the patient is susceptible to a long interval between the Q and T signals. For such a patient, the T wave is likely to appear rather late after the R wave throughout the entire cardiac cycle. With reference to the analysis using a refractory period and a CT interval, as shown in Figures 3A and 3B, a "long QT" patient may experience the T wave after the end of the CT interval. This can cause the detection circuits to detect a threshold crossing due to the variation of the T wave after the period of time CT ends in one or more detection vectors.

p00059If the “long QT” condition is verified, the method continues with the amplitude analysis, as indicated by the number 234. An illustrative method of the amplitude analysis 234 is further explained below with reference to Figures 8 and 9A a 9C. After the amplitude analysis in 234, the method is iterated using i = i + 1, and returns to block C 228 in Figure 7A, which finally sends the analysis to block 206. It should be noted that the variable {a} is not iterated in this case, since the QRS (a) and Noise (a) data elements have not been filled with data during amplitude analysis in 234.

p00060If the "long QT" condition is not verified in block 232, the method determines whether the Interval (i-1) and the Interval (i) are very similar in length. In an illustrative example, the intervals are considered to be very similar in length when their durations are within 50 milliseconds relative to each other, for example, 300 milliseconds is considered very similar in length to 320 milliseconds, although the exact parameters to determine Similarity may vary. If the intervals are very similar in length in 238, the method continues with the amplitude analysis, as shown with the number 234, as before.

p00061If the intervals compared in 238 are not similar, the method determines if the Interval (i) is longer than the Interval (i-1), as shown with the number 240. If so, then Peak (i) it is stored as QRS (a), as shown with the number 242. Otherwise, if the Interval (i-1) is not shorter than the Interval (i), the method decreases the variable {i} to i = i-1, as indicated by the number 244, and continues to step 242. Using any path, the method follows from step 242 to D, as shown with the number 230, which returns to figure 7A, directing the method to step 218.

p00062The detected episodes that are analyzed in Figures 7A and 7B and that are discovered to be QRS episodes are placed in a QRS register, while the episodes that are discovered to be noise are placed in a Noise register. The episodes that undergo amplitude analysis are placed in records as illustrated in Figure 8. In some embodiments, the method may be suspended if too many episodes are placed in records for amplitude analysis, although this is not necessarily the case.

p00063Figure 8 illustrates a block diagram for a method of evaluating the quality of cardiac signals. This method may be a part of the amplitude analysis 234 shown in Figure 7B. From the beginning block 250, the method receives a couple of episodes a and b, as shown by the number 252. Episode b must be placed in one of several registers, as indicated by the number 254, depending on the relative amplitudes of the episodes a and b. More particularly, if the amplitudes of the peaks a and b are more or less equal, within a range of +/- 15%, episode b is placed in the EQUAL register, as shown with the numbers 260, 262. If the peak b is greater than peak a, outside the range so that the two are equal, then, b is placed in the HIGH register, as shown with the numbers 256, 258. On the contrary, if peak a is much larger than peak b, then b is placed in the LOW register, as shown with numbers 264, 266. The process of placing in records of Figure 8 is used later in Figures 9A to 9C.

p00064Figures 9A to 9C illustrate a block diagram for a method of analyzing a heart signal. The method of Figures 9A to 9C is a method to calculate an Index for a given vector. The method involves the use of the methods of Figures 7A and 7B to analyze {n} adjacent episodes detected. These episodes, as explained above, can generate paired Noise and QRS data elements, which are separated and placed in a QRS record and a NOISE record. In addition, with the amplitude analysis of Figure 8, some episodes related to ambiguous intervals can be placed in records for HIGH, LOW and EQUAL.

p00065For illustrative purposes, the example of Figures 9A to 9C uses n = 11 as the number of episodes initially detected. In other embodiments, any suitable number of episodes can be used. The parameters used in the following example are simply illustrative of a way of performing the method, and can be modified to be suitable for other systems, specific circumstances or variables. The method of Figures 9A to 9C begins in a starting block, and determines whether 6 or more of the n = 11 detections have been placed within the QRS registers, as shown with the number 300. If not, the method continues with block X 302 in figure 9B. This condition may also be a condition in which a set of detections are considered and, if more than half of the detections are in the QRS records, the condition is met at 300.

p00066If there are 6 or more detections in the QRS register, then, the average amplitudes for the Noise and QRS records are calculated, as shown with the number 304. An illustrative method for performing this calculation is shown in block 306. The standard deviation for noise data is determined, as well as the arithmetic mean. Next, the average is calculated excluding any independent data (illustratively, data that falls outside a standard deviation of the arithmetic mean). The procedure is also repeated for QRS data. Alternatively, the method shown in block 312 can be used instead of calculating average values ​​for one

p00067or both of the noise and QRS data. In this somewhat simpler method, the highest data point and / or the lowest data point is removed from the data set and the average is calculated using the reduced data set, as shown with the number 312.

p00068The SNR is then calculated as the ratio between the average amplitude of the remaining episodes in the QRS register and the average of the amplitudes stored in the NOISE register, as shown with the number 308. Then, the INDEX is calculated using the amplitude average of the remaining episodes in the QRS record along with the SNR, as shown with the number 310.

p00069Next, a value for use in the assignment of indices based on the average QRS amplitude (QRSMed). In an illustrative example, the following general polynomial formula is used:

exp (N1 * [ln (N2 * QRSMed -N3)] 2

p00070SA = GAIN *

D1 * QRSMed -D2

p00071Where: GAIN = 64.0 N1 = -34.7222

p00072N2 = 0.2326 N3 = -0.6047

p00073D1 = 0.3008 D2 = -0.7821 It should be noted that the limits for QRSMed, in this illustrative embodiment, are 0 <QRSMed <4.0. A graph illustrating the relationship between SA and QRSMed is shown in Figure 10A.

p00074The illustrative index assignment method also includes calculating an SR value such as:

p00075SR = CR * (SNR) 2

p00076In which:

p00077if SNR: 3.5, CR = 0.1;

p00078if 3.5 <SNR: 10, CR = 1; and

p00079if SNR> 10, SR = 100

p00080It should be noted that the variable SR is given a constant value when the SNR is greater than 10, at least part to avoid the excessive contribution of large SNRs to the calculation of the final Index. A graph illustrating the relationship between SR and SNR is shown in Figure 10B.

p00081In some embodiments, the dynamic range of the analysis system for use in conjunction with the vector under consideration can be identified depending on the average value found and / or the arithmetic mean of the QRS data. For example, the method may include selecting a dynamic range for the analog-digital converter, the incoming signal amplifier, or other system component or components. In an illustrative example, a higher order polynomial formula of a variable (that is, in QRSMed) is used to find the index as follows:

p00082SA = {C1 * (QRSMed) 6 + C2 * (QRSMed) 5 + C3 * (QRSMed) 4 + C4 * (QRSMed) 3 + C5 * (QRSMed) 2 + C6 * (QRSMed) + C7}

p00083In which, if QRSMed: 2.0;

p00084C1 = 22.5718 C2 = -105.9666 C3 = 160.2345

p00085C4 = -88.9262 C5 = 29.6019 C6 = -1.2859

p00086C7 = 0.0087

p00087And, if QRSMed> 2.0;

p00088C1 = 56.5544 C2 = -1.069.9959 C3 = 8.310.0056

p00089C4 = -33.849.9682 C5 = 76.139.7271 C6 = -89.551.3405

p00090C7 = 43.035.7880

p00091The illustrative method that uses these coefficients is adapted for use in a system in which electronic devices that receive the patient's heart signals can have a first dynamic range of up to 2.0 millivolts and a second dynamic range of up to 4 millivolts. With such a system, the peak amplitudes for the QRS in the range of 1.7 to 2.0 millivolts create a probability of limitation and / or difficulty to use only one of the two dynamic intervals all the time. Similarly, peak amplitudes in the range of 3.5 to 4.0 millivolts also create a probability of limitation, making the latter a less preferred range.

p00092Figure 10C shows a graph comparing SA with QRSMed with this sixth order polynomial formula, and it can be seen that SA is given first and second peaks, with a depression in the area where QRSMed is approximately 1 , 7 to 2.0 millivolts. The above illustrates the use of a general polynomial formula (using serial functions such as the exponential and neperian logarithm) as well as the use of a non-continuous polynomial formula of order n. These formulas are simply illustrative, and other polynomial forms can also be used. In addition, as shown by the embodiment of Figure 10A, there is no need to especially configure the system for electronic input devices that have adjustable or multiple dynamic ranges, unless desired.

p00093Instead of a formula, a query table can also be used to calculate SA and SR. In an illustrative example, the following query table is used to find the INDEX:

CONSULTATION TABLE

<dl><dt>SA </dt><dd>Amplitude Range (mV) MR SNR </dd></dl>

<dl><dt>0,5</dt><dd> : 0,5 0,5 : 3 </dd></dl>

<dl><dt>5 </dt><dd> 0,5 – 0,65 1 3 – 3,5 </dd></dl>

<dl><dt>10 </dt><dd> 0,65 – 0,8 25 3,5 – 4 </dd></dl>

<dl><dt>18 </dt><dd> 0,8 – 1,0 50 4 – 5 </dd></dl>

<dl><dt>30 </dt><dd> 1,0 – 1,7 75 5 – 7,5 </dd></dl>

<dl><dt>20 </dt><dd> 1,7 – 2,0 100 > 7,5 </dd></dl>

<dl><dt>40</dt><dd> 2,0 – 3,0 </dd></dl>

<dl><dt>15</dt><dd> 3,0 – 3,5 </dd></dl>

<dl><dt>0,5</dt><dd> 3,5 – 4,0 </dd></dl>

p000955 The illustrative reference table is adapted for a system in which the device is adapted to detect with an input of LOW or HIGH amplitude, which has dynamic ranges of ± 2.0 millivolts or ± 4.0 millivolts. For this illustrative system, peak amplitudes in the range of 1.7 to 2.0 millivolts create a probability of limitation and / or difficulty to use only one of the two dynamic intervals all the time. Similarly, peak amplitudes in the range of 3.5 to 4.0 millivolts also create a probability of limitation, causing

p0009610 the latter is a less preferred range so that a lower SA factor is provided. The query table can be adapted to other systems that have different characteristics.

p00097In some embodiments, the dynamic range of the analysis system for use in conjunction with the vector under consideration can be established depending on the average value found and / or the arithmetic mean of

p00098fifteen QRS data. For example, the method may include selecting a dynamic range for the analog-digital converter, the incoming signal amplifier, or other system component or components.

p00099Next, the Index is determined from:

twenty SA * SR = Index

p00100Referring again to Figure 9A, with the Index calculated at 310, the index assignment is completed, and the method ends.

p0010125 Returning now to Figure 9B, the method chooses in block X 302, coming from Figure 9A. The method determines whether there are 6 or more (again, out of a total of 11, although these numbers may vary) detections in the EQUAL register, as indicated by the number 320. If so, the detection vector under consideration is providing Too many ambiguous results and too much noise, causing excessive detection that the system has a difficult time resolution. Therefore, the detection vector is declared bad, as shown by the number 322,

p0010230 and the index assignment method ends because no Index can be returned for the vector under consideration. The whole vectorization method would follow, at this point, the analysis with a different detection vector or, if all the vectors have been analyzed, it would proceed to select the best vector among those that have not been declared bad vectors.

p0010335 If the condition in 320 is not true, then, it is determined if there are at least 3 (of 11) detections in the QRS register, as indicated by the number 324. If not, the method follows in block Y 326 and in figure 9C. If there are 3

p00104or more detections in the QRS register, the method follows in 328, in which each one of the average for NOISE and QRS is calculated. This can be done according to one of the methods of blocks 306 or 312 of Figure 9A, although due to the data set initially reduced in consideration, the method of block 312 can be

p0010540 more useful, since the statistical analysis of block 306 is less useful.

p00106Next, the QRS average and the NOISE average are compared to each other to verify that the records contain values ​​that are separated from each other, as shown with the number 330. The use of 1.15 as a multiplier is simply illustrative of a form of this comparison. In another embodiment, instead of a

p00107Four. Five multiplier, a deviation having a stable value is used, or a formula such as Ax + B can be used, with A as multiplier and B as deviation. Other comparisons can also be made to determine the separation. If the values ​​in the QRS and NOISE records are too close to each other, the method jumps to step 334, in which a bad vector is declared and the index assignment ends.

p00108fifty If the condition is met in 330, the remaining values ​​in the IGUAL, HIGH and LOW records are analyzed to see which, if any, can be moved to one of the QRS or NOISE records, as indicated by the number 332. Data points in the EQUAL, HIGH and LOW registers are shifted to QRS and / or NOISE by the method steps shown in block 336. For each data point, j, the peak amplitude is compared to see if it can be adjusted in one or the other of the QRS or NOISE records due to its similarity in amplitude, using a margin of + / 15% to determine the similarity. The margin used may vary in other embodiments. If the data point, j, is sufficiently similar to the average of the QRS or NOISE record, it is then moved to that record. The stages that move j to the QRS register are shown at 338, 340, and the stages that move j to the NOISE record are shown at 342, 344.

p00109After each data point in the EQUAL, HIGH and LOW registers has been considered in block 336, it is determined again if there are 6 or more detections in the QRS register, as indicated by the number 346. If so, the Averages for the NOISE and QRS records are recalculated, as indicated by the number 350, for example, by the method of one of the blocks 306 or 312 in Figure 9A. Next, the SNR is determined as a sample with the number 352, and an index is calculated as shown with the number 354. Once the Index is at 354, the index assignment method ends, as noted in 356

p00110If the condition in 346 fails, and there are still less than 6 detections in the QRS register after block 338, the QRS, NOISE, HIGH, EQUAL and LOW registers can be restored to their original state before steps 336 to 344, such as It is shown with the number 348. The method then goes to block Y 326, and follows in Figure 9C.

p00111Returning to Figure 9C, the method follows from block Y 326. It is determined if there are 2 or more detections in the HIGH register, as indicated by the number 360. If it is not, then, a bad vector is declared, as It is indicated by the number 362, and the method of index assignment ends since no index can be returned for the vector under consideration.

p00112If there are at least 2 detections in the HIGH register, the method ranges from block 360 to 364, in which the average amplitudes for the data points in the HIGH and LOW registers are calculated. This can be done using a method similar to those shown in one of blocks 306 or 312 in Figure 9A or, instead, since the data set under consideration is already limited, the complete set of data points in each record It can be used to calculate averages in step 364.

p00113The next step is to try to move points from the EQUAL register into the HIGH and LOW registers, as shown by the number 366. This is done as indicated in block 368. For each data point k in the EQUAL register, the amplitude stored for said data point is compared with the average for the HIGH record, as shown with the number 370, to determine if it is similar to the average for the HIGH record, within a defined range of +/- 15%. If so, then, the data point k is shifted to the HIGH register, as shown with the number 372. Otherwise, the amplitude for k is compared with the average for the LOW register, again with a margin of + / - 15%, as shown with the number 374. If the amplitudes are similar, k is shifted to the LOW register, as indicated with the number 376. These steps are repeated for each point k in the EQUAL register.

p00114In some embodiments, the steps in block 368 are performed so that the test in 370 is carried out first, and if it fails, then the test in 374 continues. For example, in the context of the programmer code, one of the comparisons will appear first and, therefore, will be presented first. In this way, the +/- 15% margin overlap around the HIGH and LOW averages does not create ambiguity, since each peak (k) can be entered only in one register or the other. The same may be true for block 336 in Figure 9B.

p00115After block 368 is completed, the method determines if 4 or more detections are in the HIGH register, as indicated by the number 380. If not, then, a bad vector is declared as shown with the number 382 , since no index can be returned. If there are 4 or more points in the HIGH record, then, possible indices are determined, as indicated by 384.

p00116As shown in block 386, an OndaT_Mayor Index is calculated. This Possible Index assumes that the T wave will have an amplitude greater than the QRS signal. Therefore, detections in the HIGH register are assumed to represent T waves, while the detections in the LOW register represent QRS signals. Therefore, the SNR is calculated as the ratio between the average amplitude in the LOW register and the average amplitude for the HIGH register. This can be done using the complete data sets in each of the LOW register and the HIGH register, or the data sets can be reduced by methods such as shown in blocks 306 and 312 in Figure 9A. The variable “Amplitude” is set below as the average amplitude for the LOW register (again, a complete or reduced set of data can be used to calculate the average amplitude for the LOW register as well). The Possible Index is determined as before by the query table or diagram, although again, other methods could also be used, including the calculation of a polynomial formula.

p00117Then, as shown with the number 388, a QRS_Mayor Index is calculated. This Possible Index assumes that the QRS signal will be larger than the T wave. Therefore, it is assumed that the detections in the HIGH register represent QRS signals, and the detections in the LOW register represent T waves. The SNR is calculated as the ratio between the average amplitude in the HIGH register and the average amplitude in the LOW register in a similar way to that used in block 386, and the amplitude is calculated as the average amplitude for the HIGH register, using new complete or reduced data sets. The Possible Index is calculated below as before.

p00118After the creation of the Possible Indices (the QRS_Mayor INDEX and the OndaT_Mayor Index) in steps 386, 388, an indicator is established to ask the operator if "QRS> OndaT?" as indicated by the number

p00119390. If necessary, that is, if the largest available index is one of the possible indices, the operator can then provide an entry indicating which of the possible indices is correct. With the indicator set at 390, the index assignment method ends as indicated by the number 392, returning the two possible indexes with the indicator set.

p00120Figures 11A to 11B illustrate a block diagram for a method of vector signal analysis. The method assumes a system as illustrated in one of Figures 1A and 1B, with three available detection vectors: A-Rec, B-Rec and AB. The illustrative vectorization method of Figures 11A and 11B has a preference for vector AB, then for vector B-Rec, and the lowest preference for vector A-Rec. This is not necessary for the implementation of the invention, but Figures 11A and 11B illustrate how such preferences can be incorporated into a vectorization method.

p00121The method begins in block 400, in which the detection vector AB is analyzed. As shown with the number 402, the first query is whether an Index (AB) (an Index for the AB detection vector) has been calculated during the 400 analysis. If so, the method follows at 404 in which the Index ( AB) generated is compared to Level (AB), a threshold for the detection vector AB. If the Index (AB) exceeds the Level (AB), then the method selects the detection vector AB, as indicated by the number 406, which ends the vectorization method only if a detection vector has been sought. If multiple detection vectors have been searched, then the method can continue with the analysis of the other available vectors to select a second vector, if desired. In an alternative illustrative embodiment, the condition of premature exit of steps 404 and 406 can be omitted or, as a user option, disabled, as shown below in the method illustrated in Figure 14; The premature exit conditions are shown by the whole method of Figures 11A and 11B, although they can be omitted in selected embodiments. For this reason, the optional early stages 404, 422 and 442 are shown with a different blackness and line style in Figure 11A.

p00122If the Index (AB) does not exceed Level (AB), in step 404 (or if step 404 is omitted or disabled), then the method goes to block 408, in which the detection vector B is analyzed -Rec. Returning to block 402, if an Index for the detection vector AB cannot be calculated, it is determined at 410 if the detection vector AB is a bad vector, as indicated by the number 410. If so, the detection vector AB It can be marked as bad, and the method jumps forward to step 408. Otherwise, the method continues to block 412, in which the indicator is set to indicate that the operator input is needed in relation to the detection vector AB. With the complete analysis of the detection vector AB, the method continues to block 408, in which the detection vector B-Rec is analyzed.

p00123From block 408, it is determined again whether an Index (B) (an Index for the B-Rec detection vector) can be calculated, as shown with the number 420. If so, the method determines whether the Index ( B) is greater than Level (B), a threshold for the B-Rec detection vector, as shown with the number 422. If Level (B) is exceeded, the B-Rec detection vector is determined to be sufficient when it is used in detection and the method ends (if only one detection vector has been searched) by selecting the B-Rec detection vector, as indicated by the number 424. If Level (B) is not exceeded, the method continues to step 426, explained further below.

p00124If no Index (B) is calculated, the method determines whether the B-Rec detection vector is a bad vector, as indicated by the number 428. If so, the B-Rec detection vector can be marked as bad. , and the method continues again in step 426. If the B-Rec detection vector is not a bad vector at 428, the method establishes an indicator indicating that an operator input is needed to terminate the analysis of the B-Rec vector in order to degenerate an Index, as indicated by the number 430. Again, the method continues to step 426.

p00125In step 426, the A-Rec vector has to be analyzed. It is determined whether an Index (A) (the Index to detect the A-Rec vector) can be calculated, as indicated by the number 440. If so, then it is determined whether the Index (A) is greater than the Level ( A), a defined threshold for the A-Rec detection vector, as shown with the number 442. If so, then, the A-Rec detection vector is selected, as shown with the number 446. If Level (A) is not exceeded, the method goes to step 448 to find the best vector, which leads to block Z 450 in Figure 11B.

p00126If the Index (A) cannot be calculated in step 440, the method continues to step 452, in which it is determined whether the detection vector A-Rec is a bad vector. If so, the detection vector A-Rec can be marked as a bad vector and the method follows up to 448 and up to block Z 450 in Figure 11B. If the A-Rec detection vector is not a bad vector, the method continues to step 454 and sets the indicator to indicate that an operator input is needed to resolve an ambiguity with the A-Rec detection vector. The method then continues to step 448 and to block Z 450 in Figure 11B.

p00127In an illustrative example, one of the index assignment formats set forth above is used. In this illustrative example, Level (A) = Level (B) = Level (AB) = 1,750. However, it should be evident to those skilled in the art that the specific aspects of the index assignment formula can be modified, as well as the specific aspects of the threshold levels used in the analysis. More broadly, the illustrative example is one in which several vectors are analyzed to determine, in turn, if any of them can be characterized as "very well suited" for performing cardiac signal analysis and, if not, the most appropriate of the vectors is selected.

p00128In another illustrative example, the stage of establishing an index has two main phases: first, to determine if the captured signal is susceptible to an index assignment method, and second, to determine whether the captured signal indicates a good detection vector by calculating an index which takes into account identifiable characteristics of the captured signal including identifiable cardiac episodes and identifiable noise. The first phase takes into account the difficulty in implantable devices of separating heart signal noise, while the second phase takes into account whether the signal, once separated, is likely to help in an unequivocal analysis.

p00129Returning to Figure 11B, the method follows from block Z 450 to step 460, in which it is determined whether there are one or more Indexes already calculated. If so, the method continues to step 462 in which the vector with the highest Index is selected. In this way, the operator is not asked any questions if at least one vector provides an index.

p00130If the condition in step 460 fails, the method continues to step 464, in which the Possible Indices are used. From the Possible Indices, the largest possible index (LPS) is identified, as shown with number 466. Next, the operator is asked if QRS> T for the detection vector corresponding to the LPS, as indicated by The number 468. The response given by the operator can verify or reject the LPS, depending on whether the calculation resulting in the LPS corresponds to the response given by the operator.

p00131For example, if the LPS is the QRS_Mayor Index for the A-Rec detection vector, the operator will be asked if, for the A-Rec detection vector, QRS> ondaT? If the operator indicates "Yes", then the operator's response verifies the QRS_Mayor Index for the A-Rec detection vector, and therefore, the LPS would be verified. If, instead, the operator's answer is No, then the OndaT_Mayor Index for the A-Rec detection vector would be verified, and the QRS_Mayor Index would be discarded since the operator's response indicates that it is an incorrect calculation.

p00132As indicated by the number 470, the method then determines whether the LPS is verified by the operator's response to the question in 468. If so, then, the detection vector corresponding to the LPS is selected, as indicated by the number 472. Otherwise, the LPS is discarded, as indicated by number 474, and the method returns to step 466. If other possible indices remain, steps 466, 468 and 470 are repeated for one or more additional vectors until step 472 is reached. It should be noted that once a Possible Index is discarded in step 474, the “other ”Possible corresponding Index for said vector remains available, however, said“ other ”Possible Index will be very low so that, if there is another Possible Index available, a different vector is likely to offer better detection. In some embodiments, if the Possible Major Index for a vector assumes a different response to the "QRS> T" query than that given by the Operator, it is discovered that the vector is a bad vector.

p00133Figure 12 is a block diagram illustrating a method in which a primary detection vector and a secondary detection vector are selected for use in detecting and analyzing cardiac events. As shown with the number 500, the method first identifies a primary detection vector. This can be done, for example, using the methods described above.

p00134Next, the method includes identifying a secondary detection vector, as shown by the number

p00135502 This can be done in more ways than one. For example, as shown with the number 504, the method used to find the primary detection vector can be repeated in its entirety, ending said method in

p00136506 In another embodiment, the method simply follows the analysis after identifying the primary detection vector. As shown with the number 510, if all available detection vectors have already been analyzed, the method chooses the best one and, again, ends at 506.

p00137If not all vectors have been analyzed, then an additional vector is analyzed as shown with the number

p00138512. If this additional vector meets a first threshold in step 514, showing that the additional vector is well suited for cardiac signal analysis, then the additional vector is selected as the secondary vector, and the method ends at 506. In another embodiment , an indicator is established if all vectors have not been analyzed at 510, and an additional step is performed at the end of the method to compare the index for the primary vector with the index for the secondary vector. If the secondary vector has an index greater than the primary vector, their designations can be reversed.

p00139If the additional vector does not meet the first threshold in step 514, the method loops back to step

p00140510 The loop continues until a recently analyzed vector exceeds the threshold in step 514, or all vectors have been analyzed and the best one is selected in step 510. The use of primary and secondary vectors can take various forms, such as those described. in the US document UU. No. 7,392,085.

p00141Figure 13 is a block diagram for an illustrative embodiment. Starting at block 550, the method includes selecting a detection threshold. Next, a set of episodes is captured, as indicated by the number 552, using the detection threshold to define cardiac episodes detected together with a selected detection vector. The set of episodes is analyzed as shown with number 554. The analysis in 554 results in a determination in 556. The detection vector used to capture the set of episodes in 552 is determined in step 556 to be one of a suitable vector, an available vector or an inappropriate vector, as shown with the numbers 558, 562 and 570, respectively.

p00142If it is determined that the detection vector is a suitable vector as shown with the number 558, then it has been found that the detection vector has met the parameters selected to find a suitable cardiac detection vector. It is assumed that a vector that meets the selected parameters, in the illustrative method, is likely to be sufficient to provide accurate heart monitoring. In the illustrative method, an additional consideration of more detection vectors is considered unnecessary. Therefore, as indicated by the number 560, the method ends without further analysis of more detection vectors. The appropriate vector is then used for detection and analysis of cardiac episodes.

p00143If, on the other hand, it is determined that the detection vector is an available vector, as shown with the number 562, then the vector under consideration is considered a candidate for data collection, but does not meet the parameters to make unnecessary the additional consideration of more detection vectors. In addition, an available vector may be a detection vector that indicates ambiguity in its analysis. The method follows in 564, and determines if all available detection vectors have been analyzed. If so, the method continues to step 566, in which the best vector is selected. Step 566 may include sub-methods to resolve ambiguities, if present, in the available vector or vectors. The method ends below in

p00144560.

p00145If all the detection vectors have not been considered when in step 564, the method continues by considering a different detection vector, as indicated by the number 568. The method then returns to step 550 in the that a new detection threshold is selected for the "next" detection vector.

p00146Figure 14 is a block diagram for another illustrative embodiment. In the illustrative method of Figure 14, premature exit conditions are not used to terminate the vectorization analysis before the analysis of each of the available vectors. It should be noted, however, that a determination can be made as to what polarity will be analyzed for the polarities of each vector, for example, determining the polarity that results in the greatest positive displacement from an initial condition. In some systems, a rectified signal is used in the analysis, so that the incoming polarity is not particularly important.

p00147In the illustrative embodiment of Figure 14, the method starts at step 600 in which, as before, a detection threshold is identified. An adaptive or constant threshold can be selected, as desired. Next, a set of episodes is captured, as shown with the number 602. The set of episodes can be analyzed immediately, as shown with the number 604, or the analysis can wait until all the data collection is completed for Each vector or position. An optional box is shown with the number 620, and includes the steps of determining the characteristics of each vector, as indicated by the number 606. In the illustrative embodiment, the vectors are characterized as "available vectors", as shown with the number 608, or "inappropriate vectors", as shown with number 610. The determination in 606 can be made, for example, by comparing the SNR and / or the peak amplitude of the detected detections with one or more thresholds. Inappropriate vectors can be marked as such, and further / subsequent analysis (eg, calculation of an INDEX) can be omitted for such vectors. The block in 620 is optional, and can be omitted in some embodiments.

p00148Next, it is determined whether the acquisition stage 602 has been performed for all available vectors, as indicated by the number 612. If not, the method returns to step 600, as indicated by the number 614. If All vectors have been considered in step 612, the method continues to step 616, in which the best vector is selected, and the method ends, as indicated by the number 618.

p00149The stage of selecting the "best" vector can take several forms. In an illustrative embodiment, a vector is considered, for example, using an INDEX, and the highest "index" vector is selected. In embodiments in which vectors may be marked as "available vectors" or "inappropriate vectors", only vectors that are "available vectors" can be considered when selecting the "best" vector. Alternatively, individual vectors can be considered one at a time and compared with thresholds for each individual vector. This may allow preferential selection of a vector, even without the use of a premature exit condition. If no threshold is exceeded, the highest index vector can be selected. In another additional embodiment, different vectors may undergo different methods of index assignment, to preferentially identify a selected vector or to provide an analysis that is particularly suitable for individual vectors. For example, in a transvenous system, a first vector using the container electrode and an intracardiac electrode can be considered to be more likely to be subjected to noise than a second vector using two intracardiac electrodes. The calculation of an INDEX can take this into account by calculating the INDEX for the first vector differently than for the second vector.

p00150In each of the above embodiments, certain additional steps can be achieved to inform the operator (such as a doctor performing an implant procedure or providing a medical check-up of the patient) about vector selection. For example, the selected vector may be characterized based on its INDEX as "Good" or "Bad", for example. If the selected vector is "Good", then, the system can optionally provide, through an associated programmer, an indication thereof to the operator. Alternatively, no indication may be provided, since the operator's action is not needed and the patient is, by signaling, well adapted to the device. If the selected vector is "Bad" and has a relatively low INDEX, the operator can be informed through the programmer. This allows the operator to determine, for example, if the patient is a very high risk individual, or that the device should be removed and relied on a different mode or treatment device. The operator can also be pressed to perform an induction test (in which fibrillation is induced in the patient), under controlled clinical conditions, to determine if the device can accurately detect the arrhythmia, even with the Bad detection vector, and subsequently stimulate the patient to leave the arrhythmia. The operator may also be able to determine if there is a cause for the Bad index, for example, the incorrect placement of one or more electrodes or the malfunction of the device or a conductor. The indication of the qualities of the detection vectors is, therefore, another optional part of some illustrative methods, or it may be a feature of some illustrative devices and / or systems.

p00151The present invention includes certain embodiments in which the implantable device can perform the above methods in response to detected conditions, a request from an associated programmer, at intervals, or for other suitable reasons. The conditions detected that indicate the behavior of the vector analysis may include the occurrence of a certain repeated difficulty with the detection, for example, the identification of an excessive amount of double detections or a failure to detect episodes consistently. Another condition detected may be a rise in the SNR or a drop in the amplitude detected below a predetermined level or to a level that creates detection difficulties (for example, the range of 1.7 millivolts to 2.0 millivolts in a stepped sensor that has dynamic detection intervals that can be selected between 2.0 millivolts and 4.0 millivolts).

p00152These methods can be performed by means of an implantable cardiac stimulation device with a housing containing operating circuits, or with multiple housings attached to each other containing operating circuits distributed among them. The operating circuits may be adapted to perform several of the above steps and methods that use any analog and / or digital domain, or both, using appropriate components, devices and connections, including a microcontroller and associated memory, but without being limited to them.

p00153In an illustrative embodiment in which an implantable medical device system performs one or more of the above methods, if it is determined that an entry is required by the user to determine whether the largest possible index (LPS) identified corresponds to an index or at a possible index, the method may include additional steps. In particular, after the LPS is identified, it can be determined whether a programmer is currently in communication with the implantable medical device system. If so, then, telemetric circuits / devices are used to contact the programmer in order to request entry by the user. Otherwise, the implantable medical device system dismisses the LPS for later verification, once a programmer is available for communication, allowing user input. If more than one LPS is greater than the largest available Index, information indicating the next successive LPS (or several LPS) may be stored until communication with a programmer is available.

p00154Furthermore, the present invention includes embodiments in which at least certain stages of the above methods can be performed by a programmer adapted for use with an implantable medical device, the programmer being adapted to communicate (for such embodiments, communication is preferably, but not necessarily, wireless) with an implantable medical device. The programmer may comprise various components and appropriate circuits to perform the steps of the method. The programmer can direct the operation of the implantable medical device to perform the method or methods, or the programmer can use the implantable medical device to capture data from the patient and transfer the captured data to the programmer itself. In some embodiments, the implantable medical device may perform certain stages of the method (for example, episode detection), while the associated programmer may perform other stages (for example, analysis of the detected episodes).

p00155The present invention also includes embodiments in which machine-readable media encodes a set

p00156or sets of instructions to perform the above methods, using operating circuits of a device

p00157implantable medical device or, in some embodiments, using circuits in a programmer for use in conjunction with an implantable medical device.

p00158Illustrative embodiments also include devices, including implantable medical devices and programmers, which are adapted to perform one or more stages of the above methods, as well as systems comprising implantable medical devices and programmers that are adapted as systems for

p0015910 Perform any of these methods and / or stages.

p00160Those skilled in the art will recognize that the present invention can be practiced in a variety of ways other than the specific embodiments described and contemplated herein. Consequently, developments in form and detail can be made without departing from the scope of the present invention as described in

p00161fifteen the attached claims.