Monitoring the use of language of a patient for identifying potential speech and related neurological disorders
A method and associated computer system and computer program product for monitoring the use of language of a patient that may enable the identification, analysis and reporting on of the natural progression of various neurological or speech disorders, is provided. The method includes obtaining a verbal response of the patient to a stimulus, converting the verbal response into text and the generating of verbal and textual response data by analyzing the verbal response and response text, respectively. A response vector may be formed base on the verbal and textual response data, which may then be used to determine a reply to the patient. Improved accuracy and consistency of monitoring the use of language associated with neurological and speech disorders is achieved, leading to a greater rate of success in the treatment of these conditions.
1. A method for monitoring a use of language of a patient, the method comprising:
obtaining, by a processor of a computing system, a verbal response of the patient to a stimulus, the verbal response being captured as an audio file;
storing, by the processor, the audio file in an audio format;
converting, by the processor, the verbal response into a response text using a speech-to-text system, wherein the speech-to-text system converts the audio format to a text-based format and stores the verbal response in both the audio format and the text-based format;
first analyzing, by the processor, a phonetic component of the verbal response by analyzing the audio file to generate verbal response data;
second analyzing, by the processor, at least one of a semantic component and a grammatical component of the response text by performing a textual analysis of the response text to generate textual response data;
training, by the processor, the computing system using machine learning techniques to create an idealized class member;
classifying, by the processor, the response for each potential condition related to the use of language of the patient, wherein patient conditions are identified by comparing the response to the idealized class member;
third analyzing, by the processor, a motor response of the patient captured by video using a video-to-image analysis to extract and identify specific motor features for specific conditions, wherein a motor feature of the patient is correlated with acoustic features of the verbal response and linguistic features of the response text;
combining, by the processor, the response and the motor response to form an overall response; and
replying, by the processor, to the patient with a reply that is determined based on the overall response, eliciting further verbal responses from the patient, wherein the further verbal responses are used to enrich the idealized class member over time.
2. The method of claim 1 , wherein the stimulus is selected from the group consisting of: an audible component; a visual component; a hepatic component; a taste component; a smell component; a text-based component; a question; a picture; and a video.
3. The method of claim 1 , wherein the phonetic component of the verbal response comprises a pitch level.
4. The method of claim 1 , wherein the step of forming a reply to the patient comprises:
identifying, by the processor, a plurality of candidate conditions by analyzing the verbal response data and the textual response data;
determining, by the processor, a confidence score for each candidate condition based on the contents of the verbal response data and the textual response data;
identifying, by the processor, a potential reply to the verbal response based on the confidence scores; and
contextualizing, by the processor, the potential reply based on at least one of the previous replies to the patient, the previous recorded verbal responses and the stimulus.
5. The method of claim 4 , wherein a candidate condition is identified as a primary condition if a determined confidence score of the candidate condition is above a predetermined value.
6. The method of claim 4 , wherein the step of identifying the potential reply comprises:
comparing, by the processor, the verbal response to existing clinical knowledge;
comparing, by the processor, the verbal response to prior responses given by the patient;
comparing, by the processor, the verbal response to documented language discussions; or
any combination thereof; and
identifying, by the processor, a potential reply to the patient based on the comparison.
7. A computer program product, comprising a computer readable hardware storage device storing a computer readable program code, the computer readable program code comprising an algorithm that when executed by a computer processor of a computing system implements a method for monitoring a use of language of a patient, the method comprising:
obtaining, by a processor of a computing system, a verbal response of the patient to a stimulus, the verbal response being captured as an audio file;
storing, by the processor, the audio file in an audio format;
converting, by the processor, the verbal response into a response text using a speech-to-text system, wherein the speech-to-text system converts the audio format to a text-based format and stores the verbal response in both the audio format and the text-based format;
first analyzing, by the processor, a phonetic component of the verbal response by analyzing the audio file to generate verbal response data;
second analyzing, by the processor, at least one of a semantic component and a grammatical component of the response text by performing a textual analysis of the response text to generate textual response data;
training, by the processor, the computing system using machine learning techniques to create an idealized class member;
classifying, by the processor, the response for each potential condition related to the use of language of the patient, wherein patient conditions are identified by comparing the response to the idealized class member;
third analyzing, by the processor, a motor response of the patient captured by video using a video-to-image analysis to extract and identify specific motor features for specific conditions, wherein a motor feature of the patient is correlated with acoustic features of the verbal response and linguistic features of the response text;
combining, by the processor, the response and the motor response to form an overall response; and
replying, by the processor, to the patient with a reply that is determined based on the overall response, eliciting further verbal responses from the patient, wherein the further verbal responses are used to enrich the idealized class member over time.
8. The computer program product of claim 7 , wherein the phonetic component comprises a pitch level.
9. A processing system comprising at least one processor and the computer program product of claim 7 , wherein the at least one processor is configured to execute the computer program code of the computer program product.
10. A computer system comprising:
a processor;
a memory device coupled to the processor; and
a computer readable storage device coupled to the processor, wherein the storage device contains program code executable by the processor via the memory device to implement a method for monitoring a use of language by a patient, the method comprising:
obtaining, by a processor of a computing system, a verbal response of the patient to a stimulus, the verbal response being captured as an audio file;
storing, by the processor, the audio file in an audio format;
converting, by the processor, the verbal response into a response text using a speech-to-text system, wherein the speech-to-text system converts the audio format to a text-based format and stores the verbal response in both the audio format and the text-based format;
first analyzing, by the processor, a phonetic component of the verbal response by analyzing the audio file to generate verbal response data;
second analyzing, by the processor, at least one of a semantic component and a grammatical component of the response text by performing a textual analysis of the response text to generate textual response data;
training, by the processor, the computing system using machine learning techniques to create an idealized class member;
classifying, by the processor, the response for each potential condition related to the use of language of the patient, wherein patient conditions are identified by comparing the response to the idealized class member;
third analyzing, by the processor, a motor response of the patient captured by video using a video-to-image analysis to extract and identify specific motor features for specific conditions, wherein a motor feature of the patient is correlated with acoustic features of the verbal response and linguistic features of the response text;
combining, by the processor, the response and the motor response to form an overall response; and
replying, by the processor, to the patient with a reply that is determined based on the overall response, eliciting further verbal responses from the patient, wherein the further verbal responses are used to enrich the idealized class member over time.
11. The computer system of claim 10 , wherein the stimulus is selected from the group consisting of: an audible component; a visual component; a hepatic component; a taste component; a smell component; a text-based component; a question; a picture; and a video.
12. The computer system of claim 10 , further comprising a first data generation unit is configured to generate the verbal response data, further wherein the phonetic component comprises a pitch level.
13. The computer system of claim 10 , further comprising a second data generation unit is configured to generate the textual response data.
14. The computer system of claim 10 , further comprising a reply generation unit configured to:
identify a plurality of candidate conditions by analyzing the verbal response data and the textual response data;
determine a confidence score for each candidate condition based on the contents of the verbal response data and the textual response data;
identify a potential reply to the verbal response based on the confidence scores; and
contextualize the potential reply based on at least one of the previous replies to the patient, the previous recorded verbal responses and the stimulus.
15. The computer system of claim 14 , wherein the reply generation unit is further configured to identify a candidate condition as a primary condition if a determined confidence score of the candidate condition is above a predetermined value.
16. The computer system of claim 14 , wherein the reply generation unit is, for the process of identifying a potential reply, configured to:
compare the verbal response to existing clinical knowledge;
compare the verbal response to prior responses given by the patient;
compare the verbal response to documented language discussions; or
any combination thereof; and
identify a potential reply to the patient based on the comparison.