Verbal periodic screening for heart disease

1. A method for analyzing a voice sample of a subject to

determine a cardiac condition, comprising:

receiving a voice sample;

extracting at least one voice feature from the voice sample;

analyzing said voice feature to detect an effect of a blood flow on said at least one voice feature, said effect being at least one of a timing of said blood flow, a periodicity of said blood flow and a magnitude or change in magnitude of said blood flow;

determining a pathological cardiac condition based on said effect.

2. The method according to claim 1 , further comprising classifying said voice sample to identify voiced segments and concatenating said voiced segments, such that at least some segments are removed and remaining segment ends are smoothed.

3. The method according to claim 1 , wherein said analyzing said voice feature to detect an effect of said blood flow on said at least one voice feature comprises detecting an effect of said blood flow as effected by at least one cause affecting voice selected from a group consisting of:

ventricular beat; and

atrial beat

on said at least one voice feature.

4. The method according to claim 1 , wherein said determined cardiac condition comprises a cardiac condition selected from a group consisting of:

an abnormal heart rate;

an arrhythmic heart rate;

ventricle tachycardia;

ectopic beats;

premature ventricular contraction;

cardiac fibrillation; and

atrial fibrillation.

5. The method according to claim 1 , comprising determining a presence of an arrhythmic cardiac condition when identifying more than a predetermined number of occurrences of cardiac fibrillation.

6. The method according to claim 1 , wherein detecting an effect of a blood flow on said at least one voice feature comprises classifying one or more parameters of said cardiac behavior based on said detected effect, and said one or more parameters comprise a pulse wave shape.

7. The method according to claim 3 , wherein said detecting an effect comprises identifying an effect of at least one of a timing of said cause affecting voice, a periodicity of said cause affecting voice and a magnitude or change in magnitude of said cause affecting voice.

8. The method according to claim 1 , wherein said determining a pathological cardiac condition comprises;

calculating time intervals between consecutive occurrences of heart pulses;

matching said time intervals with reference time intervals obtained from a reference heart condition;

determining a probability of said matching; and

comparing said probability to a threshold for determining said pathological cardiac condition.

9. The method according to claim 1 , wherein said detecting an effect comprises identifying three occurrences of heart pulses and said three occurrences of a heart pulse are not identified in the same voiced segment.

10. The method according to claim 1 , wherein said detected effect comprises an effect of a magnitude, or a change in a magnitude of a heart pulse.

11. The method according to claim 1 , wherein:

said detecting an effect comprises calculating a distribution of a plurality of values of said at least one voice feature; and

said determining said pathological cardiac condition based on said effect comprises comparing a characterizing parameter of said distribution to a threshold.

12. The method according to claim 1 , further comprising validating said determining a cardiac condition and wherein said cardiac condition comprises a cardiac condition selected from a group consisting of:

an arrhythmic cardiac condition; and

atrial fibrillation.

13. The method according to claim 1 , wherein extracting the voice feature comprises performing cross coherence between a first segment of the voice sample and a second segment of the voice sample.

14. The method according to claim 1 , wherein extracting the voice feature comprises performing analysis of the speech signal selected from a group consisting of:

wavelet analysis;

Mel Frequency Cepstral Coefficient (MFCC) analysis;

glottal pulse analysis; and

Linear Predictive Coding (LPC) analysis.

15. The method according to claim 1 , wherein said method comprises receiving, extracting, analyzing and determining a pathological cardiac condition when said voice sample comprises a voice sample of spontaneous speech.

16. The method according to claim 1 , wherein said pathological cardiac condition comprises an irregular heartbeat.

17. A system for determining a cardiac condition in a voice sample provided by a subject, comprising:

a voice input for obtaining a digital voice sample of said voice sample;

a voice feature extractor that extracts at least one voice feature from said digital voice sample;

a voice feature processor that identifies an effect of a cardiac condition on said at least one voice feature; and

a cardiac condition classifier that determines the cardiac condition based on said effect,

wherein said system is configured to:

extract said at least one voice feature from said digital voice sample;

analyze said voice feature to identify an effect of a cardiac condition on said at least one voice feature; and

determine a pathological cardiac condition based on said effect.

18. The system according to claim 17 , wherein said voice input forms a part of at least one of a mobile phone, digital assistant and a car multimedia system.

19. The system according to claim 17 , further comprising a permanent memory, wherein:

said permanent memory stores at least one previously collected voice sample from said subject; and

said cardiac condition classifier determines said cardiac condition based on a change between said previously collected voice samples and said digital voice sample.

20. The system according to claim 17 , wherein said system is configured to extract, analyze and determine a pathological cardiac condition when said digital voice sample comprises a digital voice sample of spontaneous speech.

21. The system according to claim 17 , wherein said pathological cardiac condition comprises an irregular heartbeat.