Effects of Damage to the Integrity of the Left Dual-Stream Frontotemporal Network Mediated by the Arcuate Fasciculus and Uncinate Fasciculus on Acute/Subacute Post-Stroke Aphasia

(1) Background: To investigate the correlation between the integrity of the left dual-stream frontotemporal network mediated by the arcuate fasciculus (AF) and uncinate fasciculus (UF), and acute/subacute post-stroke aphasia (PSA). (2) Methods: Thirty-six patients were recruited and received both a language assessment and a diffusion tensor imaging (DTI) scan. Correlations between diffusion indices in the bilateral LSAF/UF and language performance assessment were analyzed with correlation analyses. Multiple linear regression analysis was also implemented to investigate the effects of the integrity of the left LSAF/UF on language performance. (3) Results: Correlation analyses showed that the diffusion indices, including mean fractional anisotropy (FA) values and the fiber number of the left LSAF rather than the left UF was significantly positively associated with language domain scores (p < 0.05). Multiple linear regression analysis revealed an independent and positive association between the mean FA value of the left LSAF and the percentage score of language subsets. In addition, no interaction effect of the integrity of the left LSAF and UF on language performance was found (p > 0.05). (4) Conclusions: The integrity of the left LSAF, but not the UF, might play important roles in supporting residual language ability in individuals with acute/subacute PSA; simultaneous disruption of the dual-stream frontotemporal network mediated by the left LSAF and UF would not result in more severe aphasia than damage to either pathway alone.


Introduction
The contemporary model of language organization has highlighted the important linguistic contribution of the left frontotemporal network.Modern views emphasize that language facility relies on functionally coordinated interactions of multiple brain regions located in the left inferior frontal, anterior, posterior, and inferior temporal lobes that constitute a left-lateralized frontotemporal network [1].Neuroimaging studies have revealed that the functional interactions between these frontal and temporal areas critically depend on the efficient transmission of information, and that this information flow is subserved by structural (sub)cortico-cortical connections via multiple white matter fiber pathways [2][3][4].
Hickok and Poeppel [5,6] proposed a dual stream model of language processing in the brain, including a dorsal and a ventral stream, which was one of the most influential Brain Sci.2023, 13, 1324 3 of 17 In the current work, we aimed to investigate the effects of disruption of the integrity of the left frontotemporal network on acute/subacute post-stroke aphasia (PSA).We focused on investigating its correlation with the integrity of the white matter fiber pathways rather than the activation of cortical regions.Based on the findings of previous literature, we selected the AF as the targeted neural tract of the dorsal stream pathways due to its direct connection to the frontal and temporal lobe; likewise, we selected the UF as the targeted neural tract of the ventral stream pathways.Specifically, the anatomical details of AF were taken into account in this work.In 2005, Catani et al. [11] outlined a three-segment anatomical model of the bilateral AFs.According to this model, the AF could be subdivided into one long projection fiber (i.e., a long segment, LSAF) and two short projection fibers (i.e., an anterior and a posterior segment).The LSAF suggested that it directly connects Wernicke's area to Broca's area, the anterior segment (ASAF) linked Broca's area with Geschwind's area (i.e., the inferior parietal lobule), while the posterior segment (PSAF) projected from Geschwind's area to Wernicke's area, as seen in Figure 1.It was reported that every AF segment might differently contribute to language functions [21][22][23].To better understand the language functions of the AF, it was recently suggested to take its anatomical details into account and not roughly treat it as a unifunctional entity when exploring the linguistic role of the AF [21,22].Hence, we evaluated the integrity of the long AF segment (the LSAF) rather than the whole AF fiber bundle.
Brain Sci.2023, 13, x FOR PEER REVIEW 3 of 17 whether these dual-stream pathways support language processing independently or interactively remains a matter of debate, and it is an important field for exploration.
In the current work, we aimed to investigate the effects of disruption of the integrity of the left frontotemporal network on acute/subacute post-stroke aphasia (PSA).We focused on investigating its correlation with the integrity of the white matter fiber pathways rather than the activation of cortical regions.Based on the findings of previous literature, we selected the AF as the targeted neural tract of the dorsal stream pathways due to its direct connection to the frontal and temporal lobe; likewise, we selected the UF as the targeted neural tract of the ventral stream pathways.Specifically, the anatomical details of AF were taken into account in this work.In 2005, Catani et al. [11] outlined a threesegment anatomical model of the bilateral AFs.According to this model, the AF could be subdivided into one long projection fiber (i.e., a long segment, LSAF) and two short projection fibers (i.e., an anterior and a posterior segment).The LSAF suggested that it directly connects Wernicke's area to Broca's area, the anterior segment (ASAF) linked Broca's area with Geschwind's area (i.e., the inferior parietal lobule), while the posterior segment (PSAF) projected from Geschwind's area to Wernicke's area, as seen in Figure 1.It was reported that every AF segment might differently contribute to language functions [21][22][23].To better understand the language functions of the AF, it was recently suggested to take its anatomical details into account and not roughly treat it as a unifunctional entity when exploring the linguistic role of the AF [21,22].Hence, we evaluated the integrity of the long AF segment (the LSAF) rather than the whole AF fiber bundle.We assumed that both the AF and UF were involved in language processing and hypothesized that disruption of the dual-stream frontotemporal network mediated by the left LSAF and UF would have an important impact on language abilities in individuals Brain Sci.2023, 13, 1324 4 of 17 with acute/subacute PSA, and that simultaneous disconnection of the left LSAF and UF would affect language ability significantly more than damage to either pathway alone.

Participants
Thirty-six participants (27 males/9 females, age range: 25-78 years, mean age: 53.31 ± 12.37 years; mean duration of education: 10.78 ± 3.48 years; mean time since stroke: 36.56 ± 28.25 days) with acute/subacute PSA were recruited according to the following inclusion criteria: first-ever left hemispheric stroke (i.e., infarction or hemorrhage) with normal consciousness, time post-stroke less than 90 days, age between 18-80 years, right-handedness and native Chinese speakers, the consistent presence of various types of aphasia (aphasia quotient value < 93.8), collaborating to complete a language faculties assessment and magnetic resonance imaging (MRI) scan.Those who met one of the following criteria would be excluded: history of stroke, right or bi-hemispheric stroke, duration of stroke more than 90 days, left-handedness, concomitant neuropsychiatric diseases, other aphasic syndromes such as primary progress aphasia, and being unable to finish language faculties assessment or MRI scan.
This study protocol was approved by the Institutional Review Board of our hospital, and informed consent was obtained from one family member of every participant.

Language Assessment
All patients underwent language assessment with the Chinese version of the Western Aphasia Test (WAB) after admission [24,25].The language domains, including spontaneous speech (i.e., information content and verbal fluency), auditory verbal comprehension (i.e., yes/no questions, word recognition, and sequential commands), repetition (i.e., word, phrase, and sentence repetition), and naming (i.e., object naming, word fluency, sentence completion, and responsive speech) were used to assess the patients' language ability, and the percentage score of each sub-item was calculated for analysis according to an equation: percentage score = (patient's score/total score of each sub-item) × 100%.A high percentage score of each sub-item usually represents relatively good language performance.In addition, the global aphasia severity of every patient was determined based on the aphasia quotient (AQ) value of the WAB.According to the grading criteria [26], the aphasia severity was classified into four levels: very severe (AQ, 0-25), severe (AQ, 26-50), moderate (AQ, 51-75), and mild (AQ, ≥76).The language assessment of all participants was finished by one speech-language therapist.
All original T1-weighted and diffusion image data were pre-processed with the FM-RIB Software Library (FSL version 5.0.9 software package, http://fsl.fmrib.ox.ac.uk/fsl/ fslwiki/FSL, accessed on 10 October 2022) in four steps [27].In the first step, all DICOM files were converted into NIfTI format files by means of the "dcm2niigui" tool.In the second step, head motion and eddy currents were corrected with the "eddy_correct" function.
In the third step, skull-stripping and removal of non-brain tissue was performed using the "BET v2.1" function.In the fourth step, the diffusion tensor model was linearly fitted to the diffusion-weighted images (DWI) by applying the "DTIFIT" tool.Then, fractional anisotropy (FA) images were generated for each patient.After that, both the processed T1-weighted and FA images of every patient were spatially normalized into the standard MNI152 atlas space by using the "FLIRT" tool within FSL.

DTI Tractography
The Diffusion Toolkit (version 0.6.4.1, http://trackvis.org/(accessed on 25 October 2022), Martinos Center for Biomedical Imaging, Massachusetts General Hospital) was used for the reconstruction of whole-brain diffusion tensors, and TrackVis (version 0.6.1,http://trackvis.org/(accessed on 25 October 2022), Martinos Center for Biomedical Imaging, Massachusetts General Hospital) software was applied to manually delineate the region of interest (ROI) and the visualization fiber tracking on the normalized FA images [28].Multiple ROIs approaches were applied to accomplish the reconstruction of the target fiber tracts based on a deterministic fiber-tracking algorithm (fiber assignment by continuous tracking, namely FACT algorithm).Fiber tracking was initiated with a minimum FA value at 0.20 and an angle threshold of 35 • .
To reconstruct the LSAF, a combination of three ROIs approaches were used to accomplish virtual dissection of the bilateral AFs: a frontal ROI (ROI 1) placed on the coronal slice at the entrance to the frontal lobe (anterior to the central sulcus), a temporal ROI (ROI 2) delineated in the axial slice at the entrance to the temporal lobe (below the Sylvian fissure), and a parietal ROI (ROI 3) placed tangentially to the inferior parietal cortex [21,22].The streamlines going through both ROI 1 and ROI 3 but not ROI 2 were defined as the LSAF [22,23].
To reconstruct the UF, the first ROI was placed on the temporal stem and anterior floor of the external/extrema capsules, and the second ROI was located in the anterior white matter of the temporal lobe [29,30].

Evaluation of the White Matter Pathways
Diffusion indices including the mean FA value and fiber number were used to quantitatively evaluate the structural integrity of target neural tracts.To identify the anatomical trajectory and spatial location, the reconstructed LSAF, UF, and stroke lesions were visualized on a standard T1-weighted image atlas (the MNI152 atlas).

Statistical Analyses
All statistical analyses were performed with the IBM SPSS software package (version 25.0).The Shapiro-Wilk test was applied to evaluate the normality of the data.The Pearson correlation analyses were used for parametric variables, while the Spearman correlation analyses were used for non-parametric variables.In addition, multiple linear regression analysis was also implemented to investigate the associations between language performance and the integrity of the LSAF/UF.Demographic and stroke-related variables, including sex, age, educational level, time post of stroke, and lesion volume, were selected as confounding variables.The variance inflation factor (VIF, >5) was used to diagnose multi-collinearity between variables in the multiple linear regression models.A p value less than 0.05 was considered statistically significant.

General Demographic and Clinical Characteristics
All thirty-six patients completed the language abilities assessment and MRI scan.All participants were native Chinese speakers.Twenty-one patients suffered from ischemic strokes (infarction), and others suffered from intracerebral hemorrhages (ICH).There were twenty-four patients with non-fluent aphasia, including Broca's aphasia (7 cases), transcortical motor aphasia (TMA, 3 cases), mixed transcortical aphasia (MTA, 1 case), and global aphasia (13 cases), and twelve patients with fluent aphasia, including Wernicke's aphasia (3 cases), anomic aphasia (4 cases), conduction aphasia (3 cases), and transcortical sensory aphasia (TSA, 2 cases).The aphasia severity level of all the patients ranged from mild to very severe, according to the criteria [26].The general demographic and clinical data of all participants are presented in Table 1.

Correlation Analyses between Language Assessment and MRI Measures
First, we analyzed the correlations between language performance assessment and demographic and stroke-related variables.Pearson correlation analyses were applied to analyze the correlation between age, AQ, and scores of spontaneous speech, respectively; and Spearman correlation analyses were used to estimate the correlations between education, time post of stroke, and language performance, including comprehension, repetition, naming, and speech fluency.There was no significant correlation between both demographic and stroke-related variables and language performance (p > 0.05).
Then, we analyzed the correlations between language performance assessment and MRI metrics including lesion volume and diffusion indices of the bilateral LSAF/UF with Spearman correlation analyses due to non-normally distributed variables.The results showed that lesion volume had a moderately negative relation to AQ (ρ = −0.359,p = 0.032) and the score of naming (ρ = −0.351,p = 0.036); the mean FA values and fiber number of the left LSAF had moderately to strongly positive correlations with scores of any language subsets, while the mean FA values of the right LSAF/UF had moderately positive associations with scores of some language subsets.However, no significant correlation was found between diffusion indices of the left UF and language measures, as seen in Table 2.We further evaluated the partial correlation coefficients between diffusion indices of the left LSAF and the language domain scores while removing the effects of stroke lesion volume.When lesion volume was set as a controlled variable, partial correlation analyses revealed that the diffusion indices of the left LSAF still had moderately to strongly positive correlations with all language domain scores (p < 0.05), as seen in Figure 2.

Multiple Linear Regression Analysis
We further estimated the correlations between the language abilities assessment and diffusion indices of the left LSAF/UF using multiple linear regression analysis.The results revealed an independent and positive association between the mean FA value of the left LSAF and scores of language subsets.The fiber number of the left LSAF was mainly positively related with the AQ value (B = 0.088, standard Beta = 0.636, p = 0.036, F = 1.767) and scores of comprehension (B = 0.010, standard Beta = 0.767, p = 0.013, F = 1.829).However, the diffusion indices of the left UF were not significantly associated with language performance.No significant correlations were found between the language assessment and any other demographic or clinical metrics (p > 0.05).Furthermore, the results showed no interaction effects of the integrity of the left LSAF and UF on language performance (p > 0.05).Collinearity analyses indicated no co-linearity among the variables (VIF less than 5 for all multiple linear regression models), as seen in Table 3.

Lesion Overlay Map
The lesion volumes of all participants: range = 1.44-203.50cm 3 , mean (SD) = 49.70 (44.59) cm 3 .The lesion overlay map demonstrated a heterogeneous distribution of stroke lesion sites among the participants, as seen in Figure 3.

Lesion Overlay Map
The lesion volumes of all participants: range = 1.44-203.50cm 3 , mean (SD) = 49.70 (44.59) cm 3 .The lesion overlay map demonstrated a heterogeneous distribution of stroke lesion sites among the participants, as seen in Figure 3.

Discussion
In this work, we investigated the correlation between the integrity of the left dorsal and ventral frontotemporal network mediated by the LSAF and UF, respectively, and the language performance of patients with acute/subacute PSA.Our results suggested the potential importance of the left LSAF but not the UF in supporting residual language ability in individuals with acute/subacute PSA.Meanwhile, no interaction effect of damage to the left LSAF and UF on PSA was found in our patient cohort.
Since Catani and colleagues [11] proposed the three-segment model of the AF anatomy details, the importance of evaluating the linguistic roles of tract segments rather than whole tracts has been recognized, and increasing studies reported distinct language functions of the AF segments [21][22][23].Catani et al. [11] speculated that the LSAF might be involved in phonological processing (e.g., speech repetition).López-Barroso et al. [31] supported the contribution of the left LSAF to novel word learning, and Gullick et al. [32] reported the role of the LSAF in predicting reading ability.

Discussion
In this work, we investigated the correlation between the integrity of the left dorsal and ventral frontotemporal network mediated by the LSAF and UF, respectively, and the language performance of patients with acute/subacute PSA.Our results suggested the potential importance of the left LSAF but not the UF in supporting residual language ability in individuals with acute/subacute PSA.Meanwhile, no interaction effect of damage to the left LSAF and UF on PSA was found in our patient cohort.
Since Catani and colleagues [11] proposed the three-segment model of the AF anatomy details, the importance of evaluating the linguistic roles of tract segments rather than whole tracts has been recognized, and increasing studies reported distinct language functions of the AF segments [21][22][23].Catani et al. [11] speculated that the LSAF might be involved in phonological processing (e.g., speech repetition).López-Barroso et al. [31] supported the contribution of the left LSAF to novel word learning, and Gullick et al. [32] reported the role of the LSAF in predicting reading ability.
In studies on PSA, Ivanova et al. [21] found a potential correlation between the microstructural integrity of the left LSAF, and auditory comprehension and naming ability in individuals with chronic PSA.We recently reported the correlation between the left LSAF and aphasia severity, auditory comprehension, and naming in acute/subacute PSA [22].In this study, the results of correlation analyses showed moderate to strong positive correlations between the integrity of the left LSAF and language domains, including aphasia severity, comprehension, and naming, of language assessment in populations with acute/subacute PSA, which were further determined with multiple linear regression analysis.Indeed, it was reported that the left LSAF might be involved in the pragmatic integration and higher cognitive function processes of language, as well as in the processing of syntactically complex sentences [33,34].Thus, disruption of the integrity of the left LSAF could cause deficits in the processing of complex syntactic structures, which results in impairments in the comprehension of noncanonical sentences [35][36][37].
Notably, our results also revealed a significant positive correlation between the integrity of the left LSAF and spontaneous speech, repetition, and fluency.Gajardo-Vidal et al. [38] recently pointed out that damages to both the left ASAF and LSAF were likely to result in long-lasting speech production impairments after a lesion involved in Broca's area.Kernbach et al. [39] suggested that both the ASAF and LSAF might play roles in predicting the prognosis of speech fluency.However, we could not come to conclusions about the general linguistic roles of the left LSAF in spontaneous speech, repetition, and fluency; based on the findings of previous literature, it was suggested that spontaneous speech disturbances and speech fluency impairment might be mainly associated with damage to the ASAF [40][41][42] while repetition deficits were related to damage to the PSAF [21,23,43,44].We speculated that the direct connection between the left frontal and temporal lobes, which were important for successful speech production, repetition, and speech fluency, might explain the correlation between LSAF and speech, repetition, and fluency in our study.
It was suggested that the UF might have a dual functionality: one related to a broad range of behaviors, including anxiety, psychopathy, schizophrenia, frontotemporal dementia, and autism spectrum disorder [45,46]; and the other to lexico-semantic processing [33,47].The association between the UF and language processing may be speculations regarding its anatomical position and cortical terminations [48].The anterior temporal cortex (especially the temporal pole) formed the caudal termination of the UF and was regarded as a potential 'hub' of a distributed semantic memory network [49], while the inferior frontal lobe (especially the orbitofrontal cortex) was suggested to involve encoding and processing names [9,50].Both the anterior temporal lobes and portions of the frontal lobes have been proposed to encode, store, and retrieve semantic knowledge [51,52].Therefore, the UF (mainly the left UF) might support broad language domains, including semantic processing [47], fluency and object naming [53,54], speech production deficits [55], and syntactic processes [56].
However, several studies reported no significant correlation between the integrity of the UF and language processing.For example, Breier et al. [57] found that damage to the UF was not associated with deficits in confrontation naming.Van Hees et al. [58] found no structural changes in the bilateral UF following naming treatment in chronic aphasia, and Ivanova et al. [42] suggested no effects of damage to the UF on language processing in chronic PSA.Additionally, the roles of the left AF and UF have been jointly investigated in several studies on PSA.Marchina et al. [59] reported that damage to the AF, but not to the UF, was significantly associated with impairments in spontaneous speech, including speech rate, informativeness, overall efficiency, fluency, and naming skills.Coincidentally, Hope et al. [54] also found that speech fluency and naming accuracy were significantly negatively related to the lesion load of the left AF but not of the UF in chronic PSA.Rosso et al. [60] found that the global severity of chronic aphasia was related to a critical white matter crossroad area overlapping with the left AF and the IFOF, but not with the left UF.Consistent with these previous findings, our results also supported the important influence of damage to the left LSAF rather than the UF on PSA.Differently from these studies, we took into account the sub-components of the left AF and investigated the linguistic involvement of the left LSAF while these prior studies explored the language functions of the whole AF entity.
Based on previous findings, several possible factors should be considered to account for the controversy on the linguistic role of the UF.Firstly, the UF anatomically connects regions located at the ventral and medial frontal lobe (which are not commonly associ-ated with linguistic function) rather than the classic frontal language areas (i.e., Broca's area/inferior frontal gyrus), to the anterior temporal lobe and surrounding structures [45].Second, the UF was suggested to comprise multiple sub-components.For example, Ebeling et al. [61] proposed a three-segment anatomical model of the UF, and Hau et al. [62] reported five subcomponents of the UF.Each subcomponent had different asymmetries and distinct functions.However, the linguistic roles of the subcomponents of the UF were rarely reported.Due to the limitation of DTI tractography methodology, we could not subdivide the UF and investigate the correlation between the lateralization of the UF subcomponents and linguistic functions.
Additionally, the dorsal stream (i.e., the AF) was suggested to be more strongly left lateralized while the ventral stream pathways, including the UF, were bilaterally represented [2,6].Therefore, the left AF is more likely to be destroyed by left hemispheric lesions, while the linguistic roles of the left UF may be more easily compensated by the contralateral UF after left hemisphere lesions [59,63].Catani et al. [64] reported an extreme left lateralization of the direct long segment in ~60% of the right-handed population, and found that the extreme left lateralization of the direct long segment was significantly associated with worse performance on a complex verbal memory task.They speculated that asymmetry of the direct segment might represent a key anatomical substrate for language lateralization [52,64].According to the method of Catani et al. (2007), we evaluated the lateralization of the LSAF and UF in intra-individuals with an intact left LSAF (patient ID: 1-17, Table 1) and UF (patient ID: 1-8 and 18-27, Table 1).As a result, we found a leftward distribution of the fiber number index of the LSAF (88.24%, mean ratio = 0.07 ± 0.082) and a bilateral distribution of the fiber number index of the UF (mean ratio = 0.00 ± 0.096).Our results were consistent with the previous reports [29,58,64,65], which support the potential linguistic role of the left AF rather than the UF in acute/subacute PSA; the lateralization property of the LSAF and UF might be one of the explanations for our findings.
Finally, the UF was postulated to act as a subsidiary role with the IFOF [66,67].Duffau et al. [63,68] suggested that the "semantic ventral stream" might be constituted by two parallel pathways: a direct pathway, including the IFOF, which directly connected the posterior temporal cortex to the inferior frontal lobe and was crucial for semantic processing; and an indirect pathway, consisting of the ILF linking the posterior temporal cortex with the anterior temporal lobe (ATL) and the UF, projecting from the inferior frontal lobe to the ATL, which is not systematically indispensable and could be compensated during stimulation and after surgical resection [69], as seen in Figure 4.In the study of Rosso et al. [60], the global severity of chronic aphasia was found to be related to a combined disconnection both of the left AF and the IFOF instead of the UF.Therefore, they suggested that simultaneous disruption of both the dorsal and ventral streams might provide explanations for the strong correlation with aphasia severity in chronic stroke patients.To elucidate the impacts of the IFOF on the linguistic role of the UF, we attempted to measure the integrity of both the left IFOF and UF (the reconstruction methods of the IFOF referred to in references [29,30]).However, we could not distinguish herein the effects of damage to the left UF and IFOF on language performance due to a high co-occurring injury rate of the UF and IFOF in our patients' cohort, as shown in Figure 5. Therefore, further exploration of the functional relationship between the UF and IFOF is warranted.Several limitations should be mentioned in this study.First, the sample of each grou subject was relatively small, which limits the generalizability of the results.Second, w did not evaluate another neural tract, namely the extreme capsule (EmC), which was su gested to be an important ventral pathway important for semantic processing [44,70].T EmC was proposed to be an association fiber that projected to the frontal, parietal, an temporal cortices, and thus might be related to language processing [70][71][72]; howev   Several limitations should be mentioned in this study.First, the sample of each group subject was relatively small, which limits the generalizability of the results.Second, we did not evaluate another neural tract, namely the extreme capsule (EmC), which was suggested to be an important ventral pathway important for semantic processing [44,70].The EmC was proposed to be an association fiber that projected to the frontal, parietal, and temporal cortices, and thus might be related to language processing [70][71][72]; however, Several limitations should be mentioned in this study.First, the sample of each group subject was relatively small, which limits the generalizability of the results.Second, we did not evaluate another neural tract, namely the extreme capsule (EmC), which was suggested to be an important ventral pathway important for semantic processing [44,70].The EmC was proposed to be an association fiber that projected to the frontal, parietal, and temporal cortices, and thus might be related to language processing [70][71][72]; however, given the trajectory of the EmC and the AF, a stroke lesion commonly caused co-occurring damage to these two fiber bundles and made it difficult to distinguish the linguistic effects of damage to the left LSAF and EmC in our patient sample.Third, we used a relatively rough measurement of language abilities, and did not further distinguish language dimensions (i.e., phonetics and semantics).Fourth, there is a disadvantage in the deterministic fibertracking algorithm regarding crossed fiber tracking, which may influence the reconstruction of the LSAF, UF, and IFOF.Finally, the UF is more difficult to measure successfully given the sensitivity of the diffusion sequence to artifacts around the sinuses [58], which can be observed in our patients with the heterogeneity of the UF reconstructed by DTI tractography.All these limitations need to be overcome with an enlarged sample and more advanced imaging technologies in future studies.

Conclusions
Exploring the neural substrate supporting language processing has been a longstanding topic in the field of brain neuroscience.Together with previous research, this work further expanded our knowledge about the left dual-stream frontotemporal network mediated by the LSAF and UF, and advanced our understanding of the linguistic roles of the left AF and UF.Our results suggested the potential importance of the structural integrity of the left LSAF rather than the UF in supporting residual language abilities in populations with acute/subacute PSA.Meanwhile, the current results did not suggest a functional interaction between the dorsal and ventral frontotemporal network mediated by the LSAF and UF, respectively.We discussed the possible explanations for these results, which are warranted to be further verified and generalized in the future.Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.Written informed consent has been obtained from one family member of every participant to publish this paper.

Figure 1 .
Figure 1.The virtual three-segment anatomical model of the AF: the green stream-anterior segment, the yellow stream-posterior segment, and the red stream-long segment.L, left.Figure 1.The virtual three-segment anatomical model of the AF: the green stream-anterior segment, the yellow stream-posterior segment, and the red stream-long segment.L, left.

Figure 1 .
Figure 1.The virtual three-segment anatomical model of the AF: the green stream-anterior segment, the yellow stream-posterior segment, and the red stream-long segment.L, left.Figure 1.The virtual three-segment anatomical model of the AF: the green stream-anterior segment, the yellow stream-posterior segment, and the red stream-long segment.L, left.

Figure 2 .
Figure 2. The partial correlation analyses between diffusion indices of the left LSAF/UF and language domain scores (controlled variable: lesion volume).The values, including FA, fiber, comprehension, repetition, naming, and fluency, were converted to a normal distribution by SQRT.**, p < 0.01.

Figure 2 .
Figure 2. The partial correlation analyses between diffusion indices of the left LSAF/UF and language domain scores (controlled variable: lesion volume).The values, including FA, fiber, comprehension, repetition, naming, and fluency, were converted to a normal distribution by SQRT.**, p < 0.01.

Figure 3 .
Figure 3. Lesion overlay map for all participants (N = 36).The hot color represents the regions with a high lesion overlap, while the cold color represents the regions with a lower lesion overlap.

Figure 3 .
Figure 3. Lesion overlay map for all participants (N = 36).The hot color represents the regions with a high lesion overlap, while the cold color represents the regions with a lower lesion overlap.

Figure 4 .
Figure 4.The proposed two ventral stream pathways supporting semantic language processin IFOF-the green stream, ILF-the blue stream, and UF-the red stream.

Figure 5 .
Figure 5.The reconstruction of the bilateral IFOF (the green stream) and UF (the red stream).(a) sample of the intact IFOF and UF, 15 cases; (b), a representative of damage to the IFOF, 4 cases; ( damage to the UF, 4 cases; and (d), damage to both the IFOF and UF, 13 cases.The light blue bloc are stroke lesions.

Figure 4 .
Figure 4.The proposed two ventral stream pathways supporting semantic language processing.IFOF-the green stream, ILF-the blue stream, and UF-the red stream.

Figure 4 .
Figure 4.The proposed two ventral stream pathways supporting semantic language processing.IFOF-the green stream, ILF-the blue stream, and UF-the red stream.

Figure 5 .
Figure 5.The reconstruction of the bilateral IFOF (the green stream) and UF (the red stream).(a), a sample of the intact IFOF and UF, 15 cases; (b), a representative of damage to the IFOF, 4 cases; (c), damage to the UF, 4 cases; and (d), damage to both the IFOF and UF, 13 cases.The light blue blocks are stroke lesions.

Figure 5 .
Figure 5.The reconstruction of the bilateral IFOF (the green stream) and UF (the red stream).(a), a sample of the intact IFOF and UF, 15 cases; (b), a representative of damage to the IFOF, 4 cases; (c), damage to the UF, 4 cases; and (d), damage to both the IFOF and UF, 13 cases.The light blue blocks are stroke lesions.

Author Contributions:
Conceptualization, Q.Y.; methodology, Q.Y.; software, Q.Y.; validation, Q.Y.; formal analysis, Q.Y.; investigation, Q.Y., Y.J., Y.S., X.J., T.Y. and N.L.; resources, Q.Y.; data curation, Q.Y.; writing-original draft preparation, Q.Y.; writing-review and editing, Q.Y.; visualization, Q.Y.; supervision, S.Q. and K.L.; project administration, Q.Y.; funding acquisition, Q.Y.All authors have read and agreed to the published version of the manuscript.Funding: This research was funded by the Suzhou Municipal Health Commission, grant number KJXW2022037, and the APC was funded by the Suzhou Municipal Health Commission.Institutional Review Board Statement:The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Suzhou Municipal Hospital (K-2022-072-H01).

Table 1 .
General demographic and clinical characteristics of all participants.

Table 2 .
The correlation analyses between language performance assessments and MRI measures.
Note: FA_LSAF_L, means FA of the left LSAF; FA_LSAF_R, means FA of the right LSAF; Fiber_LSAF_L, fiber number of the left LSAF; Fiber_LSAF_R, fiber number of the right LSAF; FA_UF_L, mean FA of the left UF; FA_UF_R, mean FA of the right UF; Fiber_UF_L, fiber number of the left UF; Fiber_UF_R, fiber number of the right UF; *, p < 0.05; **, p < 0.01; ×, p > 0.05.