Adaptive artificial intelligence system for identifying behaviors associated with mental illness and modifying treatment plans based on emergent recognition of aberrant reactions
One or more embodiments described herein relate to predicting, using adaptive artificial intelligence techniques, typical and aberrant physiological reactions of a patient to psychiatric counseling. Treatment plans can be determined and calculated based on previously-gathered demographic and/or biometric data, and/or modifications to treatment plans can be determined and/or implemented based on emergent recognition of reaction types, such as reclassifying reactions that would previously have been deemed typical as aberrant (or vice versa).
1. A non-transitory, processor-readable medium storing code configured to be executed by a processor to cause the processor to:
monitor biometric data of a patient collected during a counseling session administered by a first counselor, the biometric data including data representative of at least one of eye movement, body perspective, body language, facial expression, word selection, sentence structure, pauses in speech, length of a response to a question, or timeliness of the response to the question;
identify an aberrant reaction of the patient to the counseling session by applying a machine learning task to the biometric data, the aberrant reaction identified when the machine learning task detects a deviation in the biometric data from typical values of biometric data for a cohort to which the patient belongs, the machine learning task identifying the aberrant reaction without comparing the biometric data to predefined thresholds;
define a modified treatment plan for at least one of the patient, the first counselor, a second counselor, or a review board based on the aberrant reaction being identified; and
update the machine learning task with the aberrant reaction such that repeated occurrences of the aberrant reaction will cause the machine learning task to relabel the aberrant reaction as a typical reaction.
2. The non-transitory, processor-readable medium of claim 1 , the code further comprising code to cause the processor to receive the biometric data from at least one of a video camera or an audio recorder that recorded the counseling session.
3. The non-transitory, processor-readable medium of claim 1 , wherein the biometric data includes data representative of at least two of the eye movement, the body perspective, the body language, the facial expression, the word selection, sentence structure, the pauses in speech, the length of the response to the question, or timeliness of the response to the question.
4. The non-transitory, processor-readable medium of claim 1 , wherein the aberrant reaction does not correspond to a predefined type of aberrant reaction.
5. The non-transitory, processor-readable medium of claim 1 , wherein the biometric data on which the machine learning task is performed and based on which the aberrant reaction is identified does not correspond to a pattern of biometric data previously identified as corresponding to an aberrant reaction.
6. The non-transitory, processor-readable medium of claim 1 , wherein the first counselor is remote from the patient during the counseling session.
7. The non-transitory, processor-readable medium of claim 1 , wherein:
the first counselor is remote from the patient during the counseling session; and
the biometric data is collected using at least one of a video camera or an audio recorder that is local with the patient.
8. The non-transitory, processor-readable medium of claim 1 , the code further comprising code to cause the processor to:
send a signal to cause an output device to warn the at least one of the first counselor, the second counselor, or the review board of the aberrant reaction.
9. The non-transitory, processor-readable medium of claim 1 , wherein the code to cause the processor to monitor the biometric data includes code to cause the processor to monitor the biometric data in real time, during the counseling session.
10. The non-transitory, processor-readable medium of claim 1 , the code further comprising code to cause the processor to:
send a signal to cause an output device to warn the at least one of the first counselor, the second counselor, or the review board of the aberrant reaction in real time during the counseling session.
11. A non-transitory, processor-readable medium storing code configured to be executed by a processor to cause the processor to:
monitor biometric data of a patient collected during a first counseling session administered by a first counselor, the biometric data including data representative of at least one of eye movement, body perspective, body language, facial expression, word selection, sentence structure, pauses in speech, length of a response to a question, or timeliness of the response to the question;
identify an aberrant reaction of the patient to the first counseling session by applying a machine learning task the biometric data, the aberrant reaction identified when the machine learning task detects a deviation in the biometric data from typical values of biometric data for at least one of the patient or a cohort to which the patient belongs, the machine learning task identifying the aberrant reaction without comparing the biometric data to predefined thresholds; and
define a modified treatment plan including having a second counseling session performed by a second counselor based on the aberrant reaction being identified.
12. A non-transitory, processor-readable medium storing code configured to be executed by a processor to cause the processor to:
train a machine learning task using biometric data for a plurality of previous patients and an indication of treatment outcome for each previous patient from the plurality of previous patients, the plurality of previous patients belonging to a cohort that includes a patient;
monitor biometric data of the patient collected during a counseling session administered by a first counselor, the biometric data including data representative of at least one of eye movement, body perspective, body language, facial expression, word selection, sentence structure, pauses in speech, length of a response to a question, or timeliness of the response to the question;
identify an aberrant reaction of the patient to the counseling session by applying the machine learning task to the biometric data, the aberrant reaction identified when the machine learning task detects a deviation in the biometric data from typical values of biometric data for the cohort, the machine learning task identifying the aberrant reaction without comparing the biometric data to predefined thresholds; and
define a modified treatment plan for at least one of the patient, the first counselor, a second counselor, or a review board based on the aberrant reaction being identified.
13. The non-transitory, processor-readable medium of claim 12 , wherein the non-transitory, processor-readable memory includes a database containing biometric data for a superset of previous plurality patients, the patent is a first patient, and the biometric data is first patient biometric data, the code further comprising code to cause the processor to:
receive demographic information for the patient, the patient identified as being a member of the cohort based on the demographic information for the patient, the machine learning model trained using the biometric data for the plurality of previous patients associated with the cohort based on identifying the patient as belonging to the cohort.
14. The non-transitory, processor-readable medium of claim 12 , the code further comprising code to cause the processor to identify the patient as being a member of the cohort based on at least one of (i) the biometric data or (ii) demographic information for the patient.
15. A non-transitory, processor-readable medium storing code configured to be executed by a processor to cause the processor to:
receive video data and/or audio data of a first counseling session administered by a first counselor;
identify an aberrant reaction of a patient to the counseling session by applying a machine learning task to the video data and/or audio data, the aberrant reaction identified when the machine learning task detects a deviation in the video data and/or audio data from typical video data and/or audio data for at least one of the patient or a cohort to which the patient belongs, the machine learning task identifying the aberrant reaction without comparing the video data and/or audio data to predefined thresholds; and
define a modified treatment plan including having a second counseling session performed by a second counselor.
16. The non-transitory, processor-readable medium of claim 15 , the code further comprising code to cause the processor to:
extract biometric data from the video data and/or audio data, the code to cause the processor to identify the aberrant reaction further comprising code to cause the processor to identify the aberrant reaction based on the machine learning task detecting a deviation in the biometric data from typical biometric data for at least one of the patient or the cohort to which the patient belongs.
17. The non-transitory, processor readable medium of claim 16 , wherein the biometric data includes at least one of eye movement, body perspective, body language, facial expression, word selection, sentence structure, pauses in speech, length of a response to a question, or timeliness of the response to the question.
18. The non-transitory, processor-readable medium of claim 15 , the code further comprising code to cause the processor to:
send a signal to cause an output device to warn the at least one of the first counselor, the second counselor, or a review board of the aberrant reaction.