Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.
1.1Wrist motion and light
What ifWhat if the device records wrist motion and ambient light without recording heart rate?
Claim language“at least comprise heart rate”
A three-axis accelerometer and an ambient light sensor record activity and light exposure in 30-second epochs over 14 days. Sustained low-motion periods and daily light exposure patterns provide estimates of rest timing and shifts in the daily schedule.
1.2Optical upstroke level crossings
What ifWhat if pulse timing comes from rising threshold crossings without locating waveform peaks?
A low-pass filter smooths the optical pulse waveform, and a comparator timestamps each rising crossing of a threshold that tracks the baseline. A 250-millisecond lockout suppresses repeat crossings, and differences between accepted timestamps form the pulse interval series.
1.4Sleep timing from diaries
What ifWhat if continuous sleep timing measurements from a diary replace the questionnaire-derived sleep rhythm type?
Claim language“an individual's MEQ type”
The diary records sleep onset and final awakening for 14 days and marks each entry as a workday or a free day. The system encodes the average free-day sleep midpoint as sine and cosine values and appends them and the night-to-night spread in sleep onset to the heart rate variability features.
1.5Fixed thresholds for warnings
What ifWhat if fixed decision rules generate a rhythm warning without a trained evaluation model?
Claim language“a trained sleep rhythm state evaluation model”
The system tracks the clock time of the daily minimum in median heart rate after smoothing the daily profile over two-hour windows. It issues a warning when that time shifts by more than two hours from the person's baseline on at least four of the last seven days.
1.8Multiscale interval fluctuation slope
What ifWhat if a detrended fluctuation slope replaces sample entropy as a descriptor of beat interval dynamics?
Claim language“sample entropy SampE”
The processor subtracts the mean from the beat intervals, integrates the result, and removes a fitted linear trend within windows of 4 to 16 beats. It fits a line to the logarithms of window size and root-mean-square fluctuation and passes the slope to the evaluation model.
Section 2 · machine-generated
Variations
Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.
Other field of use
Sleep rhythm monitoring for maritime crew
A shipboard occupational health station collects multiday single-lead electrocardiogram recordings and a morningness–eveningness questionnaire from each crew member. Software aligns the recordings to shipboard time, detects beats, and computes time-domain, frequency-domain, and nonlinear heart rate variability features across valid recording windows. A random forest trained on labeled crew recordings combines those features with questionnaire type and displays an individual risk category at the health station.
Component substitution
Regularized logistic regression for risk classification
A multinomial logistic regression model with L2 regularization takes the place of the random forest. The training pipeline imputes missing heart rate variability measurements, standardizes continuous features, and represents questionnaire type with binary indicators. At inference, the pipeline applies the stored preprocessing parameters and feeds the resulting vector into the fitted model, which reports the risk category with the highest softmax probability.
Parameter or scale variation
Fourteen day recording with overlapping windows
A recorder captures a single-lead electrocardiogram at 250 samples per second over 14 days. Software detects beats and divides the interval series into 10-minute windows with a 5-minute step, retaining windows with at least 90 percent valid intervals. The pipeline computes time-domain, frequency-domain, and nonlinear features per retained window and aggregates each feature by its median and interquartile range. A random forest trained on vectors assembled with these same settings combines the summaries with questionnaire type to assign a risk category.
Combination with existing systems
Sleep diary and risk report integration
An electronic sleep diary records bedtime, final awakening, and questionnaire type under the same participant identifier used by the physiological recorder. A nightly service joins diary entries to beat intervals by participant identifier and recording timestamp, then computes heart rate variability features within each recorded bed-to-wake interval. The trained classifier combines the resulting summaries with questionnaire type, and the diary displays the risk category alongside the dates used in the assessment.
Risk scoring with a home sleep monitor
A home sleep monitor records airflow and oxygen saturation on a clock synchronized with a multiday electrocardiogram recorder. Software aligns respiratory event markers and oxygen desaturation markers with the beat interval series, then computes heart rate variability summaries for windows without marked events and counts the marked events per recording hour. A classifier trained on the combined feature vector uses those summaries, event counts, and questionnaire type to produce an individual sleep rhythm risk category.
Automated or data-driven variant
Automated quality checks for beat intervals
A signal quality module divides the electrocardiogram into short segments and scores each segment using clipping, baseline drift, and similarity to a median beat template. After peak detection, software flags implausible intervals and excludes windows with excessive flagged beats from feature extraction. The classifier runs when the retained windows meet a configured coverage threshold, and the interface reports insufficient data when coverage falls below that threshold.
Specific version of a part
Single lead chest recording
Two adhesive chest electrodes feed a differential amplifier and a 250-hertz analog-to-digital converter. The recorder stores timestamped electrocardiogram samples and derives a heart rate series from the intervals between successive ventricular depolarizations.
Five minute variability windows
The processor divides the beat interval series into five-minute windows and rejects windows with more than five percent flagged intervals. For each accepted window, it computes the root mean square of successive interval differences, spectral power from 0.15 to 0.40 hertz, and Poincaré plot dispersion perpendicular to the identity line.
Median imputation and scaling
The processor encodes morning, evening, and intermediate types as three indicator columns and appends them to the heart rate variability columns. It fills missing continuous values with the corresponding training column median and centers and scales continuous columns using the training means and standard deviations.
Section 3 · quoted from USPTO records
Other work
Passages reproduced word for word from two earlier-published U.S. patent documents, placed beside the language of claim 1. No characterization is offered. Each passage is marked A or B for the document it comes from; ¶ numbers are the paragraph numbers in the USPTO’s published full text of that document.
1.1acquiring physiological signals of a target user, wherein the physiological signals at least comprise heart rate;
The next operation in FIG. 4, indicated by the box numbered 402, represents the first operation of normal diagnostic processing, in which the data processing apparatus 102 (FIG. 1) collects heart rate and/or IBI data from the animal 104 and produces heart rate variability data.
A · US 2008/0004539 A1 · ¶ 52
1.2preprocessing the physiological signals and performing peak extraction thereon;
Moreover, the ECG wave can be used for recognition of interbeat intervals via R-wave detection, in accordance with processing for species differences in physiology.
A · US 2008/0004539 A1 · ¶ 63
1.3extracting multi-domain heart rate variability features from the preprocessed physiological signals by using a multi-domain feature extraction algorithm based on the processed signals;
In accordance with the invention, the HRV data analysis includes (but is not limited to) determining autonomic dynamics of the animal under study, determining standard deviation data for variation between heart beats of the received heart rate variability data (SDNN), determining Total Power, determining high frequency (HF) and low frequency (LF) components of the received heart rate variability data, and determining a plot of the HF and LF components and identifying quadrant
A · US 2008/0004539 A1 · ¶ 55
1.4constructing a multi-domain feature set by combining the extracted multi-domain heart rate variability features with an individual's MEQ type;
The initial feature fusion model is configured to: splice the training semantic feature and the training global feature to obtain a spliced training feature, calculate an autocorrelation feature based on the spliced training feature to obtain an autocorrelation feature, calculate a self-attention weight based on the autocorrelation feature to obtain a self-attention weight, and adjust the spliced training feature through the self-attention weight to obtain a fused training fe
B · US 2023/0401833 A1 · ¶ 31
1.5inputting the multi-domain feature set into a trained sleep rhythm state evaluation model, and outputting an individualized sleep rhythm disorder risk category or warning,
In order to improve the accuracy of classification and recognition, different characterizations of the sample are usually recognized and then fused to obtain fused features. Classification and recognition are performed by using the fused features.
B · US 2023/0401833 A1 · ¶ 3
1.6wherein the multi-domain heart rate variability features comprise HRV time-domain features, HRV frequency-domain features, and HRV nonlinear features, and the HRV nonlinear features comprise heart rate asymmetry features, Poincaré plot features, and a plurality of entropy complexities;
In accordance with the invention, the HRV data analysis includes (but is not limited to) determining autonomic dynamics of the animal under study, determining standard deviation data for variation between heart beats of the received heart rate variability data (SDNN), determining Total Power, determining high frequency (HF) and low frequency (LF) components of the received heart rate variability data, and determining a plot of the HF and LF components and identifying quadrant
A · US 2008/0004539 A1 · ¶ 55
1.7wherein the Poincaré plot features comprise a Poincaré plot semi-major axis SD1, a semi-minor axis SD2, and a ratio SD1/SD2 of the semi-major axis to the semi-minor axis;
1.8wherein the plurality of entropy complexities comprise approximate entropy ApEn, sample entropy SampE, fuzzy entropy FuzzyEn, fuzzy measure entropy FuzzyMEn, and distribution entropy DistEn;
1.9wherein the heart rate asymmetry features comprise a Porta index, a Guzik index, a slope index, and an area index;
1.10wherein constructing the multi-domain feature set comprises:
The initial feature fusion model is configured to: splice the training semantic feature and the training global feature to obtain a spliced training feature, calculate an autocorrelation feature based on the spliced training feature to obtain an autocorrelation feature, calculate a self-attention weight based on the autocorrelation feature to obtain a self-attention weight, and adjust the spliced training feature through the self-attention weight to obtain a fused training fe
B · US 2023/0401833 A1 · ¶ 31
1.11combining the extracted multi-domain heart rate variability features with the individual's MEQ type to construct a multi-domain feature set D, and preprocessing the feature set, wherein preprocessing of the feature set comprises processing missing values and normalizing features;
1.12wherein the individual's MEQ type is a sleep rhythm type, comprising: a morning type, an evening type, and an intermediate type; and wherein the trained sleep rhythm state evaluation model is constructed by a random forest algorithm, and constructing the trained sleep rhythm state evaluation model comprises: selecting the Gini index as a splitting criterion for constructing a decision tree, wherein the Gini index Gini(D, a) of each feature of a dataset is calculated as follows: Gini(D,a)=sum_v(|Dv|/|D|)Gini(Dv) Gini(Dv)=1−sum_{k=1}{circumflex over ( )}{n}p_k{circumflex over ( )}2 wherein p_k represents a probability of different rhythm disorder risk categories in the dataset, n represents the number of rhythm disorder risk categories, Gini(Dv) represents the Gini coefficient under each feature, a represents a feature category, and Dv represents a feature set split based on feature a; starting from a root node, calculating Gini indices of all features, and selecting a feature having a minimum Gini index and a corresponding value thereof as an optimal feature and an optimal threshold; and dividing the dataset into two subsets based on the selected optimal feature and optimal threshold, and recursively repeating the above steps for each subset by selecting another optimal feature and a threshold thereof for further splitting until a stopping condition is satisfied, wherein the stopping condition comprises: all samples belonging to a same category, no available feature for further splitting, reaching a preset maximum tree depth, or the number of samples in a subset being less than a preset threshold.
An empty cell means only that no passage was reproduced there.
Statements in these documents
Sophisticated laboratory instrumentation systems can collect very precise data streams, but typically they require multiple electrodes and even animal restraints, and the animal being studied may be so upset or hampered by the instrumentation that the collected data is affected and is not reliable.
A · US 2008/0004539 A1 · ¶ 11
Thus, it is important to collect accurate, reliable data of the correct type, to properly analyze the collected data, to draw the right conclusions from the data, and to do so without great inconvenience and without great imposition on the animal and animal owner.
A · US 2008/0004539 A1 · ¶ 12
In order to improve the accuracy of classification and recognition, different characterizations of the sample are usually recognized and then fused to obtain fused features.
B · US 2023/0401833 A1 · ¶ 3
However, splicing different characterizations directly may lead to a problem of redundancy of information among the different characterizations, causing the obtained sample feature to be inaccurate.
B · US 2023/0401833 A1 · ¶ 3
Cite aspatent.fyi, US 2026/0354751 A1, claim 1, first published 2026-11-19T03:31:31-05:00, https://patent.fyi/US20260354751/, SHA-256 fc48620ab7bfc09825450b85ebe5fc4b6f717539959feb583bd7a1a52b75e0e6