Table 1.

Basic characteristics and main findings of studies including patients with ischemic 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
Ali (2012) [46]	Ischemic Cross-sectional; 21 patients; Age 54.8 ± 8.61.	0.9–3 days; 1.5 T, 16 directions; ROI analysis and tractography, FA values, rFA.	CST	NIHSS within 1 month.	Decreased FA in affected areas compared to unaffected areas. In patients with rFA under 0.8 at admission, motor function showed poor recovery at day of discharge. The reduction in the FA values of the affected side was correlated with the degree of pyramidal tract involvements that were significantly correlated with the motor outcome on patient’s discharge day.
Kwon (2012) [56]	Corona radiata infarct; Cross-sectional; 71 patients; Age 56.01 ± 10.50.	The early scanning group (ES group) within 14 days and the late scanning group (LS group) 15–28 days; 1.5 T, 32 directions; ROI-based manual reconstruction of CST, grouping according to CST integrity.	CST	MI at onset and at 6 months.	Predictability of DTT for motor outcome was better in patients who were scanned later (15–28 days after onset) than in patients who were scanned earlier (1–14 days after onset).
Puig (2013) [44]	MCA ischemic stroke; Longitudinal 70 patients; Age 72 ± 12.	≤12 h, 3 days, and 30 days, 1.5 T, 15 directions;	CST	NIHSS, MI at 2 years.	Mean FA in affected CST at the pons on day 30 decreased progressively in line with increasing motor deficit at 2-year follow-up. Mean FA values in unaffected CST increased in line with increasing motor deficit. rFA values on day 30 decreased in line with motor deficit at 2 years. rFA on day 30 was the only independent predictor of long-term motor outcome.
Forkel (2014) [29]	First-ever left MCA infarct; Cross-sectional; 16 patients; age 63.39 ± 18.44.	Overall mean 10 ± 6 days; 3 T, 30 directions; Volume of left and right longitudinal segment, anterior segment, and posterior segment of the arcuate fasciculus.	Perisylvian language networks (long-segment, anterior segment, and posterior segment of the AF).	WAB 14 days, 6 months.	In the left hemisphere the only independent predictor of longitudinal aphasia was the lesion size. For the right hemisphere, age and volume of the long segment of the AF were predictors of longitudinal aphasia severity. Age, gender, and lesion size had an overall predictive power of 28% for longitudinal aphasia severity. Age, gender, lesion size, and volume of the right long segment of the AF had an overall predictive power of 57% for longitudinal aphasia recovery.
Groisser (2014) [62]	Ischemic MCA territory; Longitudinal; 10 patients; Age 52.6 ± 13.48; 12 healthy adults.	3 to 7 days, 1 to 2 mo, and 6 to 7 mo; 3 T; Difference between the ipsilesional CST relative to the contralesional CST in FA (ΔFA), AD (ΔAD), and RD (ΔRD) (positive values: stroke-induced increases of the ipsilesional relative to contralesional; negative values: stroke-induced decreases).	CST	Upper-limb section of the MI, NHPT 3 to 7 days (S1 acute), 1 to 2 months (S2, subacute), and 6 to 7 months (S3, chronic).	Significant loss in CST ΔFA and ΔAD at S1. Patients showed significant decreases in ΔFA and increases in ΔAD και ΔRD over time. S1 potential predictors (CST diffusion, respective motor function, and DWI lesion volume) and S2 motor outcomes: Acute loss in CST AD is a strong predictor of subacute (S2) grip and overall strength of the paretic upper limb. S1 potential predictors (CST diffusion, respective motor function, and DWI lesion volume) and S3 motor outcomes: the prognostic value of acute loss in CST AD (ΔAD) extends from the subacute to the chronic post-stroke period, and from gross to fine motor functions. S2 potential predictors (CST diffusion, respective motor function, and DWI lesion volume) and S3 motor outcomes: while subacute loss in CST FA does not improve prognosis of chronic grip and MI, it is a better prognostic indicator of chronic dexterity than the behavioral measure of dexterity.
Maraka (2014) [57]	Ischemic; Longitudinal; 23 patients; Mean age 66.7 ± 12.	3–7 days, 30 days, 90 days; 3 T, 55 directions; Multi-ROI approach with ROIs in corona radiata; PLIC, and CP bilaterally; number of virtual fibers of the ipsilesional (FNi) and contralesional (FNc) CST, fiber number ratio (FNr, FNi/FNc).	CST	UE-FM, motor items of the [mNIHSS] 3–7 days, 30 days, and 90 days.	Positive correlations between the FNr (FNi/FNc CST) and the UE-FM score at each phase of ischemic stroke. Negative correlations between FNr and mNIHSS at each phase of ischemic stroke. In both cases, the correlations became stronger from acute to subacute and chronic phases. The combination of acute NIHSS and FNr significantly predicted chronic UE-FM score.
Rong (2014) [22]	Medulla infarct; Longitudinal; 3 patients; Age 55, 54, 74.	Within days 7, 14, and 30; 3 T; FA and rFA (ipsilateral/contralateral to the infarct) for three ROIs placed along the pyramidal tract pathway; tractography of the CST using two ROIs	Medulla, CP, internal capsule, and CST.	FM, BI at each visit.	Patients 1 and 2 showed good motor recovery after 14 days, and the FA values of their affected CST were slightly decreased. The affected CST passed along periinfarct areas and tract integrity was preserved in the medulla. Patient 3 had the most obvious decrease in FA values along the affected CST, with motor deficits of the right upper extremity after 30 days. The affected CST passed through the infarct and was disrupted in the medulla.
Takenobu (2014) [28]	Ischemic; Longitudinal; 10 patients; Age 72.7 ± 6.4.	Within 2 weeks; and at 1 and 3 months; 3 T; TBSS and ROI analysis, FA.	TBSS with ROI analysis for significant clusters found in the TBSS.	FM within 2 weeks, and at 1 and 3 months.	Significantly increased FA in the red nucleus and dorsal pons in the ipsilesional side at 3 months. Significantly decreased FA in the ipsilesional internal capsule at all time points, and in the CP, corona radiata, and CC at 3 months. FA values of clusters in the red nucleus, dorsal pons, midbody of CC, and CB were positively correlated with recovery of motor function.
Feng (2015) [51]	Ischemic; Cross-sectional; 76 patients; Age 56.5 ± 14.8; 12 controls.	Acute phase; 3 T, 30 directions; Lesion load of the CST, probabilistic tractography on healthy controls, weighted CST-lesion load (wCST-LL).	CST	UE-FM Scale 2–7 days after stroke and 3 months.	Correlation between wCST-LL and motor impairment at 3 months measured by UE-FM scale. In the severely impaired subgroup (defined as UE-FM ≤10 at baseline), wCST-LL correlated with outcomes significantly better than clinical assessment.
Liu (2015) [25]	Ischemic subcortical (first infarction); Longitudinal; 18 patients; Age median 47.5; 18 controls; Age median 49 years.	1, 4, 12 weeks; 3 T, 64 directions; FA maps, longitudinal changes in CST after stroke, voxel-based correlation analysis between FA changes and FM scores over time.	CST	FM, NIHSS	Four weeks after stroke onset, FA values in the CST located in the ipsilesional PMA and pons of patients were significantly decreased compared to controls and to FA values obtained 1-week post-stroke. Twelve weeks after stroke onset, significantly decreased FA values in the WM of the ipsilesional SMA of patients compared to the values 4 weeks post-onset. No significant FA changes in the contralesional CST between patients and controls during the entire follow-up period. Significantly positive correlations between FA values and FM scores within 12 weeks after subcortical infarction were found contralesionally in the medial frontal gyrus and thalamocortical connections. Twelve weeks post-stroke, FA values for the contralesional MFG and contralesional thalamocortical connections were significantly increased compared to FA values of controls and those of one week post-stroke. FM scores were positively correlated with FA values in the MFG and thalamocortical connections across the three time points after stroke.
Moulton (2015) [52]	Thrombolysed ischemic stroke of carotid artery territory; Cross-sectional; 28 patients; Age mean 68.5 ± 16.	24 h; 3 T, 30 directions; rFA, rMD, rAD, and rRD.	Subcortical WM of PrCG, corona radiata, PLIC, CP in ipsilesional and contralesional hemisphere; CCg as control region.	NIHSS day 1, 7, and mRS at 3 months.	Decreased MD and AD in corona radiata, PLIC, and M1 and decreased FA in corona radiata and M1 (affected vs. non-affected). Lower corona radiata rAD and corona radiata rMD were associated with a poorer outcome. The corona radiata rAD was the strongest independent predictor of the clinical scores (NIHSS on day 7 and mRS at 3 months) after adjusting for total infarct volume. Significant differences in corona radiata rAD between patients with complete recanalization and those without.
Zhang (2015) [60]	Pontine infarct; Longitudinal; 17 patients; Age mean 58.3; 17 controls.	Within 7, 14, 30, 90, and 180 days; 3 T, 64 directions; ROI analysis and tractography, FA and rFA (ipsilateral/contralateral).	ROIs: medulla, CP, internal capsule, CSO; tractography: CST.	FM, mRS, and BI.	rFA in the CST above the pons decreased significantly compared with those in the contralateral side and those in control subjects within 7 days, on day 14, and on day 30 after onset. rFA above the pons on day 14 correlated positively with the FM scores on day 90 and day 180 and correlated negatively with the mRSscore on day 90 and day 180. Follow-up tractography showed regeneration and reorganization of the motor pathways.
Bigourdan (2016) [24]	Minor-to-severe supratentorial cerebral infarct; Longitudinal; 117 patients; Age: 67 (58–78).	24–72 h, 1 year; 3 T, 16 directions; Deterministic tractography, semiautomatic calculation of stroke volumes, total N of fibers ipsilateral to stroke normalized by the total N of fibers from the contralateral side, iFNr and fFNr (fiber N ratio).	CST	FMA score at 1 year.	The iFNr measured at 24 to 72 h was strongly correlated to the fFNr measured at 1 year. iFNr correlated strongly with motor recovery (improving prediction compared with using only initial FMA, age, and stroke volume).
Doughty (2016) [49]	First-time ischemic hemispheric stroke; Cross-sectional; 58 patients; Age 61.3 ± 14.2; 12 controls.	Within 80 h from onset; 1.5 T, 30 directions; Lesion maps, FA, ADC and FLAIR values, FA LI (stroke affected vs. stroke unaffected hemisphere).	CP, a stretch of the CST caudal to each stroke lesion (Nearest-5-Slice, N5S).	UE-FM assessment in the acute phase and at 3 months.	The slope of the FA LI for the N5S ROI (S-N5S) showed a weak, significant trend as a predictor of 3-mo UE-FM score. The slope of the ADC LI for the N5S ROI was also weakly predictive of 3-month outcome. Initial UE-FM and wCST lesion load were strong predictors of 3-mo outcome.
Jang (2016) [55]	Pontine infarct; Cross-sectional; 31 patients; Age 64.76 ± 10.76.	7–28 days; 1.5 T, 32 directions; Multi-ROI tractography, FA, fiber number (FN), CST area size, rFA, rFN and rCST area (affected/unaffected).	CST	MI, MBC, and FAC within 24 h and at 6 months.	Significant correlation between rFN and rCST area and all 6-month motor outcome. No significant correlation between the rFA and all 6-month motor functions.
Liu (2017) [27]	Ischemic, subcortical; Longitudinal; 50 patients; Age 24–72, median 53.5; 2 controls; Age 32–71, median 52.7.	1, 4, and 12 weeks; 3 T, 64 directions; TBSS.	Whole-brain WM analysis using TBSS.	FM	No significant changes in FA, MD and LDH in WM tracts at week 1 between the PROP group and controls. No significant differences in FA in the POOR group relative to control and PROP groups. Subjects in the POOR group had reduced MD and LDH compared to controls in ipsilesional superior corona radiata (SCR), PLIC, external capsule, posterior corona radiata, ALIC, anterior corona radiata, CP, retrolenticular part of the IC. Subjects in the POOR group showed lower MD in an ROI encompassing ipsilesional SCR, ALIC, posterior corona radiata, anterior corona radiata, and SLF compared to PROP group 1 week after stroke. Subjects in POOR group had lower LDH values in a ROI encompassing ipsilesional PLIC, SCR, and retrolenticular part of the IC compared to PROP group 1 week after stroke. Initial FMA and LDH in the ipsilesional CST in the superior corona radiata and PLIC were predictors of motor improvement within 12 weeks after stroke.
Liu (2018) [58]	Ischemic, subcortical; Longitudinal; 22 patients; Age 40–75, median 51.6; 22 controls; Age 40–71, median 50.7.	1, 4, and 12 weeks; 3 T, 64 directions; FA, MD, AD, and RD for 4 ROIs (bilateral primary motor area [PMA] and cerebral peduncle [CP]), delta diffusion metrics (d12-d initial).	Bilateral PMA and bilateral CP.	FM	AD in the ipsilesional PMA and CP was lower at W1 in both PROP and POOR groups compared to controls and values in the contralesional PMA and CP. AD in the ipsilesional PMA in both PROP and POOR groups significantly increased at W4 and W12 compared to AD at W1, but no significant changes in AD in the ipsilesional CP at W4 and W12 compared to W1. Significant decreases in FA and increases in RD and MD in WM of the ipsilesional PMA and CP in both PROP and POOR groups at W4 and/or W12 compared to W1 and those in contralesional areas. For all of the patients, ΔFMA-UE-observed was greater in patients with higher ΔAD in the ipsilesional PMA. For PROP patients, only FMA-UEii predicted ΔFMA-UE-observed. For POOR group, only lesion volume was related to ΔFMA-UE-observed.
Liu (2018) [59]	Thalamic infarct; Longitudinal; 12 patients; Age 50.33 ± 7.62; 12 controls; Age 51.47 ± 8.43.	1 week, 4 weeks, 3 months and 6 months; 3 T, 64 directions; FA and MD for each ROI (ipsilesional and contralesional side for patients, mean values for both sides for controls).	Regions in corona radiata pathway: thalamus, corona radiata, and CSO.	NIHSS, BI, and NIHSS8.	Gradual increase of FA values in ipsilesional and contralesional thalamic radiation fibers from W1 to W6. No changes in controls. Negative association between FA increases and NIHSS and NIHSS8 decrease. Positive association between FA increases and BI increases.
Etherton (2019) [61]	Ischemic Longitudinal; 42 patients; Age 70.2 ± 9.2, with (n= 22) and without (n = 20) early neurological improvement.	Within 12 h and 3–5 days post stroke; 1.5 T (acute phase), 3 T follow-up (not for DTI); WMH and chronic strokes, acute infarct volume on DWI, PWI-DWI mismatch ration, FA, MD, RD and AD in WMH and NAWM contralateral to the acute infarct (probabilistic ICBM-atlas).	WMH and NAWM contralateral to the acute infarct.	NIHSS Admission, day 3–5 post-stroke.	NAWM RD and MD values were lower in the early neurological improvement + group compared to the early neurological improvement − group. Increasing NAWM RD decreased the likelihood of early neurological improvement. NAWM MD was an independent predictor of early neurological improvement (multivariate logistic regression).
Kulesh (2019) [63]	Ischemic; Cross-sectional; 100 patients; Age 68.3 ± 11.1; 10 controls.	5–10 days; 1.5 T, 12 directions; ROI analysis and tractography, FA values, and rFA.	CST (level of PLIC, pons), GIC, ALIC, CB, SLF, IFOF, SCC, infarction and the area within 3 cm from it.	Measures on day 3, 10, and at discharge: NIHSS, Frenchay Arm Test, BBS, HAI, RMI, MoCA, FIM, and mRS.	The indices of FA of the ipsilateral upper SLF and CB, FA and the size of the infarct focus, rFA of the CB, CST and the ALIC, as well as the FA of the splenium and the knee of the intact hemisphere were the most valuable predicting factors of functional outcome of acute IS. The integrity of the associative tracts of the affected hemisphere was more valuable than the microstructure of the intact hemisphere and rFA for predicting global outcome. The tracts of the intact hemisphere were more important for the restoration of complex rehabilitation spheres, such as cognitive status and daily living and social skills, which ensure patients’ independence.
Mahmoud (2019) [37]	Ischemic; Cross-sectional; 60 patients; Age 24–75, mean 58.2.	Within 2 days; 1.5 T, 12 directions; FA of WM tracts.	3D fiber tractography with multi-ROI technique and regions drawn in the unaffected portion of the WM tracts at the side of infarction and corresponding area at the contralateral hemisphere; degree of FA reduction of WM tracts at the site of infarction [mild (0.4), moderate (0.2–0.3), severe (0.1)]; and classification of WM tracts as disrupted, displaced, and preserved.	NIHSS at admission and after 3 months.	Residual neurological deficits in patients with disrupted tracts. Near-complete clinical recovery was found in patients with non-disrupted tracts. Significant association between the degree of FA reduction in the affected tracts and the clinical score at admission and the clinical recovery after 3 months.
Moulton (2019) [18]	Thrombolysed ischemic stroke of carotid artery territory; Cross-sectional; 45 patients (17 motor, mean age 65.5, and 28 aphasia, mean age 69.6).	24 h; 3 T, 30 directions; Tractography with AD and rAD (affected/unaffected), asymmetry maps for voxel-based analysis.	Second and third branches of the SLF (SLF-II and -III, respectively) and CST as part of the motor network in addition to the left and right AF, IFOF, ILF, and UF.	NIHSS, JTT, and AHS at 3 months.	Significant differences in AD between affected and unaffected tracts with the exception of the ILF in the language cohort. rAD of the CST was the sole independent predictor in the DTI model explaining 70.1% of the variance in motor impairments at 3 mo. rAD in AF—along with age and initial aphasia severity—was an independent predictor of 3-month aphasia outcome. These tract-specific results were confirmed in voxel-based analysis without a priori assumptions.
Keser (2020) [30]	Left hemispheric ischemic stroke; Longitudinal; 24 patients; Age 36–86.	Within 2 weeks and 6–12 months; DTI (no other information are available); Reconstruction of FAT and AF, FA and RD, WM lesions, actual volumes.	AF and FAT	BNT within 2 weeks and 6–12 months.	Acute phase: no lateralization between left and right AF and FAT (FA, RD). Chronic phase: Diminished FA with time in both hemisphere which was more pronounced for left FAT compared to right. The recovery rate did not correlate with acute right and left DTI values of AF and FAT. Longitudinal FA of the right AF was a significant predictor of naming recovery.
Berndt (2021) [23]	Acute occlusions of a large intracranial vessel of the anterior circulation who underwent mechanical recanalization; Cross-sectional; 165 patients; No demographics.	Median, 3 days; interquartile range [IQR], 3–4 days; maximum, 7 days; 3 T, 15 directions; Probabilistic tractography, FA [FA index = (FAI − FAH)/(FAI + FAH); I = infarcted, H = healthy non-affected.	CST (PLIC, PED)	NIHSS, mTICI, mRs at 90 days.	FA index of PLIC showed weak negative correlations to the NIHSS at the time of MR imaging. FA index was reduced in the acute poststroke phase with a correlation to clinical presentation, especially in case of peripheral infarcts. For peripheral infarcts, a strong effect of the FA index on clinical outcome (mRS) existed.
Darwish (2021) [54]	Supratentorial ischemic stroke; Longitudinal; 30 patients; Age 61.32 ± 4.34.	Time of admission and 1 month; 1.5 T, 32 directions; FA, rFA, MD, and FN.	CST (pons)	NIHSS at admission, and after 1, 6, and 9 months.	A significant negative correlation was found in moderate and severe motor deficit between rFA and FN in the ipsilateral CST at the rostral part of pons after 1 month of infarction and NIHSS score at 6 months.
Liu (2021) [32]	Ischemic subcortical, supratentorial; Longitudinal; 33 patients; Age 53.4 ± 1.81; 33 controls; Age 53.7 ± 2.32.	1, 4, and 12 weeks; 3 T, 64 directions; ROI analysis with mean FA and MD values.	Bilateral inferior cerebellar peduncle (JHU-ICBM-DTI-81-WMPM-90p).	FM	Decreased FA and increased MD in contralesional ICP at W12 after stroke compared with controls and with values at W1. Decreased FA but not MD changes in ipsilesional ICP at W4 and W12 compared to controls. Changes in FM-LL scores of the affected limb correlated positively with FA changes and negatively with MD changes in the contralesional ICP.
Xia (2021) [33]	Ischemic stroke of internal capsule and neighboring regions; Longitudinal; 34 patients; Age mean 63.4 34 controls; Age mean 61.2.	Weeks 1, 4, and 12; 3 T, 30 directions; Tractography with multi-ROIs, FA values of CST from PLIC; rsfMRI data for FC.	CST	NIHSS, MMSE, FMA, and BI after each scan.	FA values in the ipsilesional remaining CST showed a decline during the first week and increased longitudinally over the next 12 weeks after stroke. No FA changes in contralesional CST from week 1 to week 12 after stroke. Positive correlation between percentage changes of FA in ipsilesional remaining CST and those of interhemispheric FC in patients during weeks 1 to 4 after stroke. The increase of both interhemispheric FC and ipsilesional CST-FA were significantly correlated with the greater change of FMA between weeks 1 and 4 after stroke. Only increased FA of ipsilesional CST was significantly correlated with the greater change of FMA at weeks 4–12 after stroke compared to interhemispheric FC.
Li (2022) [26]	Pontine infarction; Longitudinal; 16 patients; Age 56.9 ± 7.7; 16 controls; Age 54.3 ± 5.2.	1 week (T1), 1/2 months (T2), 1 month (T3), 3 months (T4), and 6 months (T5) after onset; 3 T, DTI 64 directions; ROI FA analyses; rs-fMRI.	ATR, CST, CCG, CH, FMAJ, FMIN, IFOF, ILF, SLF, UF and SLF-TP.	UE-FM before and after each scan.	The FA values (TBSS) were significantly lower in the pontine infraction group than in the control group at T1, tended to expand at follow-up until the completely disappeared at T5. The FA values of the ATR, CST, CCG, FMAJ, FMIN, IFOF, ILF, SLF, UF, and SLF-TP were significantly increased at T2 compared with that at T1. There were no significant differences in the FA values of any specific ROIs among T3, T4, and T5.
Shaheen (2022) [38]	MCA ischemic stroke Longitudinal; 34 patients; Age 64.4 ± 6.7; 17 controls; Age 61.8 ± 5.2.	Baseline (within 7 days?) and 6 months; 1.5 T, 12 directions; CST tractography with multi-ROIs (anterior mid-pons, CP, PLIC, and corona radiata), FA and rFA, tract categorization as intact, displaced, and disrupted.	CST	NIHSS, MRC, mRS, and MI at baseline and 6 months.	Reduced baseline FA values of the CSTs on the affected side compared to the contralateral and controls. Lower mean baseline FA lesion side and rFA compared to follow-up. Patients with high baseline FA and rFA showed good recovery response. FA on the lesion side and rFA were negatively associated with follow-up NIHSS and MRS scores and positively associated with follow-up MI scores.