# US 2021/0241116 A1 · Quantification and estimation based on digital twin output

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2021/0241116 A1
- Document ID: US20210241116A1
- Status: Published application
- Kind: A1
- Application number: 17/097,058
- Filed: 2020-11-13
- Published: 2021-08-05
- Earliest claimed: 2020-02-05
- Assignees: Koninklijke Philips N.V.
- Applicants: Koninklijke Philips N.V.
- Inventors: Murtaza Bulut; Cornelis Petrus Hendriks; Lieke Gertruda Elisabeth Cox; Valentina Lavezzo
- Web page: https://patentgrep.com/patent/US20210241116A1

## CPC

- A61B 5/4824: {Touch or pain perception evaluation (A61B5/4821 takes precedence)}
- A61B 5/7267: {involving training the classification device}
- G06F 3/015: {Input arrangements based on nervous system activity detection, e.g. brain waves [EEG] detection, electromyograms [EMG] detection, electrodermal response detection}
- G06F 18/2113: by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
- G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting
- G06F 18/217: {Validation; Performance evaluation; Active pattern learning techniques}
- G06F 18/24137: {Distances to cluster centroïds}
- G06F 30/27: using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 3/09: Supervised learning
- G06N 3/044: Recurrent networks, e.g. Hopfield networks
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

Embodiments are described herein for training machine learning models using digital twin outputs, including for purposes of pain quantification and estimation. In various embodiments, simulation parameter(s) may be applied to a digital twin created for a subject. The digital twin may simulate an anatomical structure of the subject, and the simulation parameter(s) may cause the digital twin to generate digital twin output that simulates behavior of the anatomical structure in response to the one or more simulation parameters. Data indicative of the digital twin output may be applied as input across a pain estimation machine learning model to generate pain estimation output. Based on the pain estimation output, a quantification of pain, e.g., an estimated pain score, may be rendered using an output device and/or used to select a treatment plan.

## Claims

### Claim 1

A method implemented using one or more processors, comprising:
applying one or more simulation parameters to a digital twin created for a subject,
wherein the digital twin simulates an anatomical structure of the subject, and
wherein the one or more simulation parameters cause the digital twin to generate digital twin output that simulates behavior of the anatomical structure in response to the one or more simulation parameters;
applying data indicative of the digital twin output as input across a pain estimation machine learning model to generate pain estimation output,
wherein the pain estimation machine learning model is trained to estimate a quantification of pain that the subject would experience based on application of the simulation parameters, and
wherein the pain estimation output indicates the quantification of pain; and
based on the pain estimation output, rendering the quantification of pain using an output device or selecting a treatment plan.

### Claim 2 (depends on 1)

The method of claim 1, wherein the pain estimation machine learning model comprises a convolutional neural network.

### Claim 3 (depends on 1)

The method of claim 1, wherein the anatomical structure comprises a heart of the subject, and the digital twin output comprises an electrocardiogram (“ECG”) signal.

### Claim 4 (depends on 1)

The method of claim 1, wherein the one or more simulation parameters comprise a physical or daily activity.

### Claim 5 (depends on 1)

The method of claim 1, wherein the one or more simulation parameters comprise a medication or therapy.

### Claim 6 (depends on 1)

The method of claim 1, further comprising comparing the digital twin output to one or more additional outputs of one or more healthy digital twins to determine at least one digital twin delta, wherein each of the one or more healthy digital twins simulates an anatomical structure in a different healthy subject, and wherein the data indicative of the digital twin output comprises the at least one digital twin delta.

### Claim 7 (depends on 6)

The method of claim 6, wherein the one or more additional outputs of one or more healthy twins comprise a plurality of additional outputs of a plurality of healthy digital twins, the at least one digital twin delta comprises a plurality of digital twin deltas, and the data indicative of the digital twin output comprises a value calculated based on the plurality of digital twin deltas.

### Claim 8 (depends on 7)

The method of claim 7, wherein additional output of a first healthy digital twin of the plurality of healthy digital twins is weighted more heavily than a second additional output of a second healthy digital twin of the plurality of healthy digital twins based on one or more criteria.

### Claim 9 (depends on 8)

The method of claim 8, wherein the one or more criteria include a healthy subject underlying the first healthy digital twin having a greater measure of similarity to the subject than another healthy subject underlying the second healthy digital twin.

### Claim 10

A method of training a pain estimation machine learning model, the method implemented using one or more processors and comprising:
providing a first set of digital twins to simulate an anatomical structure in a plurality of pained subjects;
determining pain scores observed by the plurality of pained subjects;
applying one or more simulation parameters to each digital twin of the first set to generate a plurality of digital twin outputs, wherein each digital twin output simulates behavior of the anatomical structure in a respective subject of the first set in response to the one or more simulation parameters;
applying data indicative of the plurality of digital twin outputs as inputs across a pain estimation machine learning model to generate pain estimation outputs; and
training the machine learning model based on comparison of the pain estimation outputs to the pain scores.

### Claim 11 (depends on 10)

The method of claim 10, further comprising:
providing a second set of digital twins to simulate the anatomical structure in a plurality of pain-free subjects; and
applying the one or more simulation parameters to each digital twin of the second set to generate additional digital twin outputs, wherein each additional digital twin output simulates behavior of the anatomical structure in a respective subject of the second set in response to the one or more simulation parameters.

### Claim 12 (depends on 11)

The method of claim 11, further comprising comparing the additional digital twin outputs to the plurality of digital twin outputs generated from the first set of digital twins to determine a plurality of digital twin deltas, wherein the data indicative of the plurality of digital twin outputs comprises the plurality of digital twin deltas.

### Claim 13 (depends on 11)

The method of claim 11, further comprising:
determining additional pain scores associated with the plurality of pain-free subjects;
applying data indicative of the additional digital twin outputs as inputs across the pain estimation machine learning model to generate additional pain estimation outputs; and
training the machine learning model based on comparison of the additional pain estimation outputs to the additional pain scores.

### Claim 14 (depends on 11)

The method of claim 11, wherein determining the additional pain scores associated with the plurality of pain-free patients includes setting each of the additional pain scores to a control value.

### Claim 15 (depends on 1)

At least one non-transitory computer-readable medium comprising instructions that, in response to execution of the instructions by one or more processors, cause the one or more processors to perform the method of claim 1.