Trainable diagnostic system and method of use

A system for analyzing functional data including a functional waveform acquired during a cardiac cycle of a subject to identify an ischemic myocardium comprising:

• a signal processor connected to receive the functional waveform of the cardiac cycle of the subject and measuring a TR interval of the cardiac cycle and normalize the functional waveform based on a resting heart rate;
• a neural network having a plurality of weights determined based on pre-selected learning functional waveforms and receiving and processing the TR interval and the normalized functional waveform based on the plurality of weights to generate at measure of a change in the TR interval when the subject is subjected to stressed condition and when the subject is subjected to unstressed conditions;
• and
• a diagnostic interpretation module receiving the measure of the change in the TR interval from the neural network and calculating a probability of the subject having ischemic myocardium based on the measure of the change in the TR interval when compared to at least one probability distribution related to individuals having ischemic myocardium;
• and
• wherein the diagnostic interpretation module classifies the functional waveform as at least one of ‘normal’, ‘abnormal’, and ‘uncertain’ based on the probability of the subject having ischemic myocardium.

A method for diagnosing a cardiac condition comprising the steps of:

• acquiring functional information from a heart of a subject during a cardiac cycle;
• normalizing the acquired functional information about a cardiac phase with respect to a resting cardiac phase;
• applying the normalized functional information to a neural network that has been trained to recognize indication of a given pathology in functional information;
• determining a probable diagnosis that the subject is afflicted with the given pathology;
• and
• displaying the probable diagnosis.