Table 2.

Basic characteristics and main findings of studies including patients with hemorrhagic stroke.

1st Author (Year)	Type of Stroke, Study Design, Participants, Age (years)	Time of DTI Acquisition, Field Strength, DTI Parameters, DTI Analysis/Metrics, Additional Imaging/Electrophysiology	Anatomical Region Examined	Outcome Scale Utilized	Main Findings
Kuzu (2012) [36]	Hemorrhagic; Longitudinal; 23 patients; Age 65.26 ± 12.70.	Within 3 days, day 14, 30, 60, and 90; 3 T, 6 directions; FA values of bilateral CP.	Bilateral CP	NIHSS at day 90.	The mean FA value on day 3 was significantly higher in the good recovery group than in the poor recovery group. The mean FA value gradually decreased until day 90 in the poor recovery group, but not in the good recovery group. The FA value and the motor function score on day 3 were independent factors for predicting the motor function outcome. FA value on day 3 could predict motor function outcome with a sensitivity of 100% and a specificity of 77.8% at an FA value of 0.7 on day 3.
Wang (2012) [64]	ICH of putamen, thalamus, with or without intraventricular bleeding; Cross-sectional; 27 patients; Age 60.4 ± 10.7.	Either within 3 days or at 2 weeks after onset; 1.5 T; Average FA and MD of 6 manual regions manually drawn unilaterally in the anterior CP, and rFA.	CP	mRS, FIM, NIHSS, and PG at 6 monhts.	The FA values within 3 days and after 2 weeks of ICH onset were significantly decreased at the affected side, but the mean MD remained unchanged. The rFA within 3 days was negatively correlated with the PG, positively correlated with the FIM scores, and negatively correlated with the mRS scores at the end of follow-up. The rFA at 2 weeks was positively correlated with the FIM and negatively correlated with mRS scores and PG at the end of follow-up. A cutoff point of 0.955 for rFA within 3 days had 53% sensitivity and 100% specificity for predicting good motor outcome over 6 months. A cutoff point of 0.875 for rFA values obtained after 2 weeks had a sensitivity of 76% and a specificity 89% for predicting good motor outcome over 6 mo.
Koyama (2013a) [43]	Thalamic and/or putaminal hemorrhage; Cross-sectional; 32 patients; Age 64.25 ± 13.43.	Days 14–18; 3 T, 12 directions; FA values (left and right ROIs), rFA.	CP and the corona radiata/internal capsule.	MRC at 1 month.	Cerebral peduncle rFA values had statistically significant relationships with MRC scores. rFA values for the corona radiata/internal capsule showed less significant relationships.
Koyama (2013b) [41]	Thalamic and putaminal hemorrhage; Cross-sectional; 12 patients; Age 62.92 ± 14.64.	14–18 days after admission; 3 T, 12 directions; FA values (left and right ROIs), rFA.	CP	BRS, FIM-motor at 3–7 months.	High statistically significant relationship between rFA and upper extremity function. Medium statistically significant relationship between rFA and lower extremity function. Analysis of rFA and FIM-motor scores did not reveal statistical significance.
Ma (2014) [35]	Hemorrhagic (basal ganglia); Longitudinal; 23 patients; Age 34–67, mean 54 ± 9.	Day 0, 30, and 90; 1.5 T, 6 directions; Seed ROI in the CST portion of the ipsilesional CP	CST	MFS on day 90.	Mean FA value on day 0 was significantly lower in the poor outcome group compared to the good outcome group. FA value gradually decreased in the poor outcome group until day 90 after onset while it continously increased in the good outcome group. The MFS obtained at day 90 after onset was significantly correlated with the initial FA value in the affected cerebral peduncle. FA value on day 0 could predict motor function outcome with a sensitivity of 88.89% and specificity of 92.86% at the initial FA value of 0.45.
Tao (2014) [50]	Hemorrhage; Cross-sectional; 32 patients with CST affected by hematoma; Age 65.59 ± 17.09; 12 controls.	4 days after ICH onset; 1.5 T, 30 directions; FA in the affected and unaffected side at nearest slices under the lesion and at CP, rFA in the CP, mean rFA of “consecutive 5 slices” under the hematoma, and rFA in 5 individual slices.	CST	mRS at follow-up visits in the outpatient clinic.	The rFA measured in the cerebral peduncle, but not the 5 slices below the lesion, was significantly lower in the group with poor functional outcome than those with good functional outcome. The ICH score had greater areas under ROC curve in predicting functional outcome compared to the mean rFA
Cheng (2015) [53]	Hemorrhage; Cross-sectional; 48 patients; Age 62 ± 14.	Within 14 days (mean 7 days); 3 T, 30 directions; FA, ADC, rFA, rADC, and CST tractography-based groups: preservation around the haematoma (type A), partial interruption (type B), and complete interruption at or around the haematoma (type C).	Pons, CP, perihaematoma oedema, and corona radiata.	MI at admission, at 1 and 3 months.	14 patients in type A, 20 patients in type B, 14 patients in type C. No significant differences on rFA between groups on pons, CP, oedema except for corona radiata. No significant differences on rADC between the three groups on four ROIs. The rFA at the corona radiata was significantly positively correlated with the MI at admission, at 1 month and 3 months. No correlation between rADC and motor functions at any time.
Koyama (2015) [42]	Putaminal and/or thalamic hemorrhage; Cross-sectional; 40 patients; Age 61.83 ± 12.85.	14–21 days after admission; 3 T, 12 directions; FA values of left and right ROIs, rFA.	CP	BRS (shoulder/elbow/ forearm, hand, lower extremity), FIM-motor, and length of total hospital stay from admission to acute medical service to discharge from long-term rehabilitation facility (LOS).	Moderate-to-high positive correlations between rFA and S/E/F BRS, hand BRS, LE BRS, and FIM-motor (though weaker). Negative moderate correlation between rFA and LOS. Logistic model fit was moderate for shoulder, elbow, forearm BRS and lower extremity BRS and much higher for hand BRS.
Fragata (2017) [31]	Acute spontaneous SAH; Cross-sectional; 60 patients; Age 59 (35–86).	<72 h; 1.5 T, 20 directions; FA and ADC averaged between left and right ROIs, whole brain FA and ADC.	Frontal CSO, parietal CSO, lentiform nucleus, thalamus, PLIC, CCg, CCs, and mid-pons cerebellar peduncles.	mRS, presence of DCI at 3 months.	FA values in the cerebellum showed an association with DCI (multivariate analysis). No association between early ADC values and occurrence of DCI (multivariate analysis). Weak evidence of an association between ADC values at the frontal centrum semiovale and mRS at 3 months.
Min (2020) [48]	Putaminal hemorrhage; Longitudinal; 12 patients; Age 50 ± 12.	Within 1 day, 3 weeks, 3 months, and 6 months; 3 T, 25 directions; CST.	CST with ROIs at the pons.	BMS, MBI, mRS, NIHSS, JHFT, and MI.	The mean FA value on pons in the affected side of the brain was significantly higher in the good outcome group compared to the poor outcome group at baseline and 6 months after initial treatment. No significant differences in mean FA value in the non-affected side between the good and the poor outcome groups at all time points. The rFA of the good outcome group was significantly higher compared to the poor outcome group at baseline and 6 months after initial treatment. FA values and rFA at baseline significantly correlated with the sum of BMS scores 6 months after initial treatment.
Gong (2021) [45]	Hemorrhage; Cross-sectional; 75 patients; Age 58.8 ± 13.9.	Around 3 weeks after stroke; 1.5 T, 15 directions; ROI-based reconstruction of CST and qualitative data for no disruption, partial disruption, complete disruption of CST.	CST connecting the hand–knob area of the PrCG and the CP.	BRS-H at post-stroke 3 weeks and 3 months.	Degrees of integrity of the CST was negatively correlated with the BRS-H at post-stroke both 3 weeks and 3 months. Patients with intact CST or completely disrupted CST shown by DTI did not show significant improvement in the BRS-H at post-stroke 3 months. Patients with partially disrupted CST showed significant improvement in the BRS-H at post-stroke 3 months compared to 3 weeks.