Organizations: College of Artificial Intelligence, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu, China · Key Laboratory of Brain-Machine Intelligence Technology, Ministry of Education, Nanjing University of Aeronautics and Astronautics, Nanjing, Jiangsu, China · State Key Laboratory of Transvascular Implantation Devices and TIDRI, School of Medicine, Zhejiang University, Hangzhou, 310009, Zhejiang, China · Zhejiang Key Laboratory of Medical Imaging Artificial Intelligence, Zhejiang University, Hangzhou, 310058, Zhejiang, China
Stroke is a disorder of the circulation, yet the haemodynamic state that immediately precedes it has rarely been observed, because continuous physiological recordings before a stroke are seldom available. Photoplethysmography (PPG), recorded by every pulse oximeter and by hundreds of millions of wearable devices, captures the peripheral pulse beat by beat, but its morphology is largely discarded. Here we ask whether pulse waveform morphology encodes haemodynamic signatures of impending stroke. Using in-hospital stroke as a natural experiment that provides continuous PPG before physician-adjudicated stroke recognition, we studied intensive-care (MIMIC-III, n=176) and emergency-department (MC-MED, n=154) cohorts and described each beat with interpretable descriptors of pulse timing, amplitude, contour and variability. At a nominal 360-min horizon, morphology separated pre-recognition from stable windows with F1 scores of 0.941 and 0.989 (sensitivity 0.998 and 0.980), with models transferred unchanged to the second cohort, and outperformed heart rate alone on identical windows. The signature was dominated by systolic crest time and its change from each patient's own baseline, feature attribution was concordant across cohorts (Spearman ρ=0.865), and morphology recorded in the first two hours of care stratified the subsequent clinical course (F1 0.714 and 0.691). In high-risk patients without stroke, 29% and 14% of windows were positive, but five-window persistence rules limited alerts to 17% and 2% of patients. The results demonstrate that the routinely recorded pulse thus carries haemodynamic imprint of the period preceding stroke, positioning pulse morphology as a candidate continuous cerebrovascular biomarker.
Figures & tables
Figure 1: PPG analysis framework. a , PPG-linked cohorts and dataset roles. Pre-recognition cohorts are counted as patients and clinical-course cohorts as waveform-linked records. Strict pre-recognition subsets are shown separately. b , Waveform processing and extraction of 17 timing, amplitude, ratio and variability features. c , ResNet-1D modelling and evaluation. Pre-recognition classification uses nominal 240-, 300- and 360-min window-construction settings relative to documented recognition. The original clinical-course model uses the first 120 min of PPG after intensive-care admission (MIMIC-III) or emergency-department arrival (MC-MED). Outcomes are improved versus worsened/deceased. The clinical-course counts in panel a describe the waveform-linked source records.
Characteristic
Warning cohort
Clinical-course cohort
MIMIC-III
MC-MED
MIMIC-III
MC-MED
Cohort size
Eligible cohort, n
176
154
1,810
3,472
Age
Mean, years
66.9 [-1pt] (15.3)
71.0 [-1pt] (15.0)
67.1 [-1pt] (14.2)
71.0 [-1pt] (16.8)
Median, years
68.8 [-1pt] [54.2–79.6]
75.0 [-1pt] [61.0–83.0]
68.0 [-1pt] [56.0–79.0]
75.0 [-1pt] [62.0–85.0]
Table 1: Demographic and clinical characteristics of the PPG-linked study cohorts.
Model
F1
Precision
Recall
a, Nominal 240-min setting
MIMIC-III (internal validation)
PPG morphology
0.7956±0.0027
0.6681±0.0041
0.9833±0.0106
CHA2DS2-VASc
0.6570±0.0250
0.5890±0.0204
0.7440±0.0388
Nwosu–EHR
0.6650±0.0314
0.5570±0.0378
0.8240±0.0344
Teoh–EHR
0.7260±0.0268
0.5730±0.0316
0.9920±0.0066
Table 2: Performance of PPG and EHR reference models for stroke warning across three temporal settings.
Figure 2: Subgroup performance. a–d , F1 of the original MC-MED ensemble by clinical risk, race, age and sex. e–h , F1 of the patient-grouped MIMIC-III ResNet-1D at probability 0.5 by hypertension, diabetes, atrial fibrillation and recent drug exposure. i,j , Diagnostic-code strata of the original 360-min warning analysis, using five-model mean probabilities at cutoff 0.5 in 174 MIMIC-III and 112 MC-MED patients from Table 1 . Both coding families and other recorded diagnoses are separate groups. Bar length and vertical marks show point estimates, and darker blocks show 95% patient-cluster bootstrap intervals (10,000 draws for a–d, 2,000 for e–j). Labels give unique patient counts. k,l , Precision in the same diagnostic groups, with 2,000 patient-cluster bootstrap intervals. Cyan denotes MIMIC-III and magenta MC-MED.
Figure 3: Paired morphology, HR and HRV comparison. a , F1 . b , Paired F1 differences; blocks show 95% intervals from 2,000 patient-cluster resamples. c , Confusion counts, coloured by proportion within each true class (13,238 warning and 1,879 stable windows). All models use the same 15,117 windows from 164 MIMIC-III patients and cutoff 0.5. Morphology uses ResNet-1D; HR/HRV use logistic regression with nested tuning.
Model
Precision
Recall
F1
MIMIC-III
PPG (ResNet-1D)
0.6486±0.0290
0.8062±0.1260
0.7139±0.0538
CHA2DS2-VASc
0.4908±0.0327
0.6690±0.0275
0.5658±0.0279
Nwosu–EHR
0.5138±0.0452
0.5120±0.0677
0.5097±0.0358
Teoh–EHR
0.5853±0.0781
0.7361±0.0821
0.6457±0.0374
Yang–EHR
0.5566±0.0728
0.6397±0.0679
0.5890±0.0229
Table 3: Performance of PPG and EHR reference models for stroke prognosis prediction.
Figure 4: Instance-level SHAP explanations in MIMIC-III. a , Pre-recognition warning. b , Clinical-course stratification. Contributions are to the positive-class logit relative to the model baseline; magenta increases the logit and cyan decreases it.
Figure 5: Cohort-level SHAP distributions, cross-cohort agreement and temporal attribution. a , SHAP beeswarm distributions for the warning model in MIMIC-III and MC-MED. Each plotted point represents an evaluated window; colour indicates the feature value. Features are ordered by cohort-specific mean absolute SHAP value. b , Agreement of mean absolute attribution across 17 features (Spearman ρ=0.865 , two-sided P=7.32×10−6 ). Each point represents one feature; the dashed line indicates equal attribution across cohorts. Both axes are expressed in units of 10−3 . c , Mean absolute attribution of the five leading warning features in six hourly bins preceding documented recognition. Axis labels indicate bin centres ( −5.5 to −0.5 h); cell values and colours share a common 10−3 scale. The matrix shows non-monotonic temporal variation in feature attribution. d , The ten highest-ranked features in the clinical-course model, ordered by mean absolute SHAP value.
Figure 6: Tsp –HR association and temporal change in stroke and clinical non-stroke controls. a,b , Within-patient centred HR and crest time; hexagon intensity denotes segment count. c , Patient-level Pearson correlations, with median and interquartile range. d,e , Hourly patient-averaged crest time and ECG HR; shading denotes pointwise 95% patient-bootstrap intervals. Stroke observations extend to −0.5 h and control observations to −2.5 h relative to their assigned reference. f , Correlation change Δrrm ( −5 to −2 h minus −8 to −5 h). Densities show 2,000 patient-bootstrap estimates, with the central 95% shaded darker and observed estimates marked by vertical ticks. Cyan denotes stroke and magenta clinical non-stroke controls. Panels a–e use 60 segments from 20 stroke patients and 101 segments from 21 controls; c uses the 11 stroke and 18 control patients with at least three variable observations. In f, the early and late periods included 5 and 7 stroke patients and 17 and 16 controls, respectively.
Setting
Control units, n
Retained windows, n
Persistent alerts, n
Window FPR
Persistent- alert rate
Stroke-window TPR
a, MIMIC-III
240 min
229
2,997
14
0.1672
0.0611
0.9833
300 min
227
2,954
33
0.2881
0.1454
0.9647
360 min
230
2,910
40
0.2938
0.1739
0.9981
b, MC-MED
240 min
165
1,257
5
0.1663
0.0303
0.9341
Table 4: False-alert burden among high-risk non-stroke controls across nominal temporal settings.
Figure 7: Persistent alerts in non-stroke controls. a,b , Alert events per 100 patient-days and time in alert for one to ten consecutive windows; outlined columns mark the five-window rule. c , Window false-positive rate and stroke-window recall (TPR). Cell values are percentages; both cohort colour scales span 0–100%. d , Proportion of controls with a five-window alert; darker blocks show 95% Wilson intervals and black horizontal marks show estimates. Counts below the axis are alerted/total control units. Cyan denotes MIMIC-III and magenta MC-MED.
Figure 8: Available-waveform alert coverage and conditional intervals to documented recognition. a,b , Patient-equal mean probabilities in MIMIC-III and MC-MED, with pointwise 95% patient-cluster bootstrap intervals. Patient support varies by bin; bins containing one patient are descriptive. c , Persistent-alert occupancy among patients contributing to each time bin, with bootstrap intervals. d , Patients with a confirmed alert: 168/174 (96.6%) and 46/112 (41.1%); whiskers are Wilson 95% intervals. All patients remain in the denominator, including 62 MC-MED patients with fewer than five windows. e , Intervals among alerted patients, expressed in hours. White points and thick bars denote medians and interquartile ranges: 5.25 [3.24–7.22] and 1.68 [0.42–2.49]. f , One MIMIC-III trace; the dashed line marks the frozen threshold 0.791 and shading identifies above-threshold output. Confirmation occurs at the end of the fifth consecutive window with probability greater than 0.791. For patients with multiple recordings or anchors, the earliest confirmed relative time is retained. Cyan circles denote MIMIC-III and magenta squares MC-MED.
Figure 9: Patient-level waveform coverage. a , Analysed duration. b , Windows per patient. The upper row shows the available-segment alert population; the lower row shows the patient-grouped population. Cyan and magenta bars show MIMIC-III and MC-MED, respectively, on a common 0–80% scale. Labels identify the largest bin in each cohort distribution. Duration merges overlapping intervals within patients. Unequal-width bins are displayed as categories.
Measure
MIMIC-III
MC-MED
a, Signal quality
Evaluated patients, n
312
158
SQI, mean (s.d.)
0.85 (0.11)
0.90 (0.08)
SQI, median [IQR]
0.87 [0.79–0.93]
0.93 [0.86–0.96]
b, Waveform availability and processing
Warning hourly retention
98.4–98.8%
82.9–98.8%
Table 5: PPG signal quality, waveform retention and processing success across datasets.
Feature
Definition and physiological description
Time-domain morphology (4 channels)
Tsp
Onset-to-peak interval; duration of the systolic upstroke.
SI
Pulse amplitude divided by crest time; mean systolic upstroke slope.
Aoff
PPG amplitude at pulse offset; end-of-pulse signal level.
Tsys/Tdia
Onset-to-peak duration divided by peak-to-offset duration; relative lengths of the rising and falling phases.
Derivative-based timing (4 channels)
Table 6: Fixed 17-channel PPG morphology input.
Figure 10: PPG feature extraction. a , Schematic processing workflow based on pyPPG [ 42 ] . b , A measured pulse, velocity (VPG), acceleration (APG) and third derivative (JPG) on a common time axis. Landmarks use their detected sample positions; coincident c,d points share a label. Onset, peak and offset are on , sp and off ; arrows span the onset-to-peak interval Tsp and onset-to-VPG-valley interval Tv . fs is sampling frequency and x the filtered PPG. Table 6 defines all 17 descriptors.
College of Artificial Intelligence, Nanjing University of Aeronautics and Astronautics, China · State Key Laboratory of Transvascular Implantation Devices and TIDRI, Hangzhou 310009, China · Zhejiang Key Laboratory of Medical Imaging Artificial Intelligence, Hangzhou 310058, China +1
Department of Computer Science, Luddy School of Informatics, Computing, and Engineering, Indiana University Bloomington, Bloomington, IN, USA · Department of Anesthesia, Indiana University School of Medicine, Indianapolis, IN, USA · Department of Intelligent Systems Engineering, Luddy School of Informatics, Computing, and Engineering, Indiana University Bloomington, Bloomington, IN, USA +1