Signal analysis of cardiac and other patient medical signals
A system and method provide detecting and monitoring cardiac electrophysiological changes by determining differences between a reference signal representing a heartbeat signal and an analysis signal representing another heartbeat signal. A set of signal data consisting of a number of heartbeat signals is acquired as a function of time. Individual heartbeat signals within the set of acquired signals are identified, an individual signal comprising data values representing a plurality of sample points during a single heartbeat. One of the identified individual signals or a derived signal representing an average of a subset of the identified individual signals is selected as an analysis signal. A reference signal is identified. The sample points of the reference signal are aligned with the sample points of the analysis signal. A single value representative of mutual correspondence between the sample points of the reference signal and the corresponding sample points of the analysis signal is calculated. Data indicating a calculated value is communicated to a destination for viewing by a user.
1. A method of analyzing cardiac electrophysiological signals, the method comprising the activities of:
acquiring a set of signal data representing a plurality of cardiac electrophysiological signals as a function of time;
identifying individual signals within the set of acquired signal data, an individual signal comprising data values representing a plurality of sample points over a single heartbeat;
selecting one of the identified individual signals or a derived signal representing an average of a subset of the identified individual signals as an analysis signal having a plurality of first sample points of said plurality of sample points;
identifying data representing a reference signal, the reference signal comprising data values representing a plurality of second sample points;
aligning the first and second sample points;
calculating using a probability function, a single value representative of differences between the aligned corresponding first and second sample points; and
communicating data indicating the calculated single value to a destination for viewing by a user.
2. A method of analyzing cardiac electrophysiological signals, the method comprising the activities of:
acquiring a set of signal data representing a plurality of cardiac electrophysiological signals as a function of time;
identifying individual signals within the set of acquired signal data, an individual signal comprising data values representing a plurality of sample points over a single heartbeat;
selecting one of the identified individual signals or a derived signal representing an average of a subset of the identified individual signals as an analysis signal having a plurality of first sample points of said plurality of sample points;
identifying data representing a reference signal, the reference signal comprising data values representing a plurality of second sample points;
aligning the first and second sample points;
calculating a single value representative of differences between the aligned corresponding first and second sample points; and
communicating data indicating the calculated single value to a destination for viewing by a user wherein
the reference and analysis signals represent cardiac electrical activity and the activity of calculating the single value is a calculation of mutual correspondence, performed according to:
MC
(
S
1
,
S
2
)
=
∑
i
=
1
N
p
i
(
s
1
i
,
s
2
i
)
/
N
wherein S 1 is the reference signal represented over a set of sample points i numbered from 1 to N, S 2 is an analysis signal represented over a set of sample points i numbered from 1 to N, and p i (s 1i ,s 2i ) is a joint probability function for data values between the corresponding sample points of S 1 and S 2 .
3. The method of claim 1 , wherein the activity of calculating the single value is a calculation of mutual correspondence, performed according to:
MC
(
S
1
,
S
2
)
=
[
∑
i
=
1
N
min_amplitude
(
s
1
i
,
s
2
i
)
max_amplitude
(
s
1
i
,
s
2
i
)
]
/
N
wherein S 1 is the reference signal represented over a set of sample points i numbered from 1 to N, S 2 is analysis signal represented over a set of sample points i numbered from 1 to N, min(s 1i ,s 2i ) is the lesser of the data values of s 1i and s 2i for each corresponding sample point i, and max(s 1i ,s 2i ) is the greater of the data values of s 1i and s 2i ; for each corresponding sample point i.
4. The method of claim 1 , wherein the activity of calculating the single value further includes the activities of:
transforming the signal data of the reference signal from data as a function of time to data as a function of frequency;
transforming the signal data of the analysis signal from data as a function of time to data as a function of frequency;
calculating a mutual correspondence between the reference signal and the analysis signal using the data as a function of frequency; and
calculating a time-frequency mutual correspondence value between the reference signal and the analysis signal, performed according to:
MC
time
-
frequency
=
1
-
A
-
B
A
+
B
wherein A is the value representative of mutual correspondence of the reference signal and the analysis signal calculated using data as a function of time, B is the value representative of mutual correspondence of the reference signal and the analysis signal signal calculated using data as a function of frequency, and ∥x∥ represents the absolute value operator.
5. The method of claim 1 , wherein the activity of calculating the single value further includes the activities of:
transforming the signal data of the reference signal from data as a function of time to signal data as a function of other variables;
transforming the signal data of the analysis signals from data as a function of time to signal data as a function of other variables;
calculating a mutual correspondence between the reference signal and the analysis signal for each set of transformed signal data; and
calculating a value representative of a multi-dimensional mutual correspondence of the reference signal and the analysis signal, performed according to:
MC
multi
-
dimension
=
1
-
∑
N
A
i
-
A
_
∑
N
A
i
where A i is the value representative of a mutual correspondence of the reference signal and the analysis signal as a function of each variable to which the reference signal and the analysis signal were transformed; Ā is the average value of the value representative of a mutual correspondence of the reference signal and the analysis signal as a function of each variable to which the reference signal and the analysis signal were transformed; and N is the number of different variables to which the signal data was transformed.
6. The method of claim 5 , wherein the set of signal data acquired represents electrocardiogram signals for a patient and a value representative of the mutual correspondence of the reference signal and the analysis signal is provided to assist in the monitoring or diagnosis of cardiac electrophysiological changes.
7. The method of claim 5 , wherein the activity of calculating a value representative of the mutual correspondence of the reference signal and the analysis signal is performed only over a selected portion of an electrocardiogram signal.
8. The method of claim 1 , wherein the reference signal is one of: (a) derived by an averaging calculation of one of, a set of signals from a patient and cardiac electrophysiological signal data from a repository.
9. The method of claim 1 , wherein the reference signal is derived from one of: (a) current electrocardiogram signals for a patient, (b) previous electrocardiogram signals for the patient, and (c) previous electrocardiogram signals for patients whose physical characteristics match the patient.
10. A system for analyzing cardiac electrophysiological signals, the system comprising:
an acquisition processor for acquiring a set of signals representing a plurality of cardiac electrophysiological signals as a function of time;
a signal processor coupled to said acquisition processor for
identifying individual signals within the set of acquired signals, each individual signal comprising data values on a plurality of sample points over a single heartbeat;
selecting one of the identified individual signals or an average of a subset of said identified individual signals to obtain an analysis signal having a plurality of first sample points of said plurality of sample points,
identifying data representing a reference signal, the reference signal comprising data values representing a plurality of second sample points, and
aligning the first and second sample points;
a calculation processor coupled to said signal processor for calculating using a probability function, a value representative of differences between the aligned corresponding first and second sample points; and
an output processor coupled to said calculation processor for outputting the calculated value to a user.
11. The system of claim 10 , wherein said set of signals acquired by said acquisition processor comprises electrocardiogram signals.
12. The system of claim 11 , wherein said signal processor averages a plurality of electrocardiogram signals to remove noise and artifacts from the acquired set of signals.
13. The system of claim 10 , wherein said calculation processor operates to calculate said value representative of differences between the reference signal and the analysis signal by calculating mutual correspondence, according to:
MC
(
S
1
,
S
2
)
=
∑
i
=
1
N
p
i
(
s
1
i
,
s
2
i
)
/
N
wherein S 1 is the reference signal represented over a set of sample points i numbered from 1 to N, S 2 is an analysis signal represented over a set of sample points i numbered from 1 to N, and p i (s 1i ,s 2i ) is a joint probability function for data values between the corresponding sample points of S 1 and S 2 .
14. The system of claim 10 , wherein said calculation processor operates to calculate said value representative of differences between aligned corresponding first and second sample points, according to:
MC
(
S
1
,
S
2
)
=
[
∑
i
=
1
N
min_amplitude
(
s
1
i
,
s
2
i
)
max_amplitude
(
s
1
i
,
s
2
i
)
]
/
N
wherein S 1 is the reference signal represented over a set of sample points i numbered from 1 to N, S 2 is analysis signal represented over a set of sample points i numbered from 1 to N, min(s 1i ,s 2i ) is the lesser of the values of s 1i and s 2i for each corresponding sample point i, and max(s 1i ,s 2i ) is the greater of the values of s 1i and s 2i for each corresponding sample point i.
15. The system of claim 10 , further comprising
a memory coupled to said acquisition processor and said signal processor for storing the acquired set of signals and, wherein
said signal processor transforms a stored set of signal data as a function of time into a set of signal data as a function of frequency and stores the set of transformed signal data in said memory.
16. The system of claim 14 , wherein said calculation processor
calculates mutual correspondence between the reference signal and the analysis signal as a function of frequency; and
calculates a time-frequency mutual correspondence between the reference signal and the analysis signal, performed according to:
MC
time
-
frequency
=
1
-
A
-
B
A
+
B
wherein A is the mutual correspondence between the reference signal and the analysis signal calculated as a function of time, B is the mutual correspondence between the reference signal and the analysis signal calculated as a function of frequency, and ∥x∥ represents an absolute value operator.
17. The system of claim 10 , further comprising
a memory coupled to said acquisition processor and said signal processor for
storing the acquired set of signals and, wherein
said signal processor transforms a stored set of signal data as a function of time into sets of signal data and transforms the signal data of the analysis signals as a function of time into sets of signal data; and
said calculation processor calculates the mutual correspondence A i between the reference signal and the analysis signal for each set of signal data and transformed signal data, to produce a value representing a multi-dimensional mutual correspondence between the reference signal and the analysis signal, according to:
MC
multi
-
dimension
=
1
-
∑
N
A
i
-
A
_
∑
N
A
i
where A i comprises a mutual correspondence for each domain; Ā is the average value of the mutual synchronicities of all domains; and N is the number of different domains to which signal data is transformed.