Discussion

In this retrospective cohort of 23 patients undergoing resection of non-contrast-enhancing IDH-mutant gliomas of the left frontal lobe, we observed distinct patterns of postoperative language impairment associated with involvement of the FAT and AF. Involvement of the FAT was predominantly associated with transient postoperative deficits, whereas persistent impairments were more frequently observed in patients with AF involvement. Patients with combined FAT and AF involvement showed the poorest recovery. Although these findings should be interpreted cautiously given the sample size and study design, they support the concept that different white matter pathways may contribute differently to postoperative language recovery.

FAT impairment – SMA syndrome?

The FAT has been shown to be involved in speech initiation and fluency (Catani et al., 2013; Catani and Mesulam, 2008). Lesion and stimulation studies have demonstrated that FAT disruption can cause transient speech arrest or initiation delays, often resolving over days to weeks (Ille et al., 2018; Young et al., 2020, 2022; Fujii et al., 2015). This evidence underpins the first school of thought, which situates the FAT within the SMA initiation network and predicts reversible deficits.

Our findings indicate that isolated FAT lesions were associated with postoperative language impairments, including initiation deficits and aphasic symptoms (Fig. 2, Fig. 3, Fig. 4). While these deficits were common immediately after surgery, they were predominantly transient and largely resolved at follow-up. This observation is consistent with previous reports showing that FAT injury can be associated with language disturbances while rarely resulting in permanent aphasia (Catani et al., 2013; La Corte et al., 2021; Vassal et al., 2014; Kinoshita et al., 2015; Chernoff et al., 2018; Zhong et al., 2022).

The transient nature of these deficits may reflect functional recovery and network reorganization, but it is also compatible with the emerging view that the FAT supports domain-general executive control processes that contribute to language in the dominant hemisphere rather than serving an exclusively language-specific function. Consistent with this perspective, recent studies have implicated the FAT in broader cognitive and action control mechanisms beyond language (Dick et al., 2019; Serrano-Sponton et al., 2024; Tagliaferri et al., 2023). Given the limited sample size and the absence of non-language control measures in the present study, our findings should therefore be interpreted as supporting an association between FAT lesions and transient postoperative language deficits, rather than establishing a language-specific role of the FAT.

AF and persistent language impairment

By contrast, the AF is a well-established component of the dorsal stream as conceptualized by Hickok and Poeppel's dual-stream model (Hickok and Poeppel, 2007, 2015). It provides the principal perisylvian link between posterior temporal regions and frontal language cortices (Catani and Mesulam, 2008; Catani et al., 2005). Consistent with prior studies, we found that AF disruption was significantly associated with (short-term and) persistent aphasia at follow-up (Salvati et al., 2021; Tuncer et al., 2021). This is in line with evidence that posterior AF segments, particularly at the superior temporal gyrus and perisylvian junction, are especially critical for lasting language function (Tuncer et al., 2021; Saur et al., 2008). Several studies suggest that these two regions within the AF are particularly vulnerable to persistent language deficits, likely due to the dense convergence of subcortical white matter pathways in these areas (Catani and Mesulam, 2008; Tuncer et al., 2021; Forkel et al., 2014; Saur et al., 2008).

Combined FAT and AF disruption

Assessing both tracts concurrently allowed us to observe that combined FAT and AF involvement was associated with higher rates of postoperative aphasia and reduced recovery. While we did not focus in this study on cases of isolated AF injury to directly contrast, our findings suggest that combined tract damage may amplify the emergence of both transient and persistent deficits. This observation contributes new evidence to the debate on tract-specific versus network-level vulnerability of language function in glioma surgery (Duffau et al., 2014; Nieberlein et al., 2023).

Clinical implications

From a surgical perspective, our findings support careful consideration of both the FAT and AF during preoperative planning and intraoperative mapping. Preservation of the FAT may contribute to reducing transient postoperative language disturbances, whereas preservation of the AF remains particularly important for lasting language outcome (Vooijs et al., 2024). Sparing the AF, particularly its posterior segments, remains paramount for maintaining (short- and) lasting language abilities. While intraoperative mapping remains the gold standard for functional preservation in language-eloquent surgery, additional advanced tractography provides a practical framework to identify these pathways and to guide resections that maximize tumor removal while safeguarding language function (Salvati et al., 2021; Ille et al., 2018; Tuncer et al., 2021; Seidel et al., 2024; Fujii et al., 2015; Intraoperative mapping and monitoring during, 2022). Studies have shown that the potential for language reorganization in glioma patients is strongly region-dependent: frontal language areas such as the left inferior frontal gyrus often exhibit high plasticity with perilesional or distributed intrahemispheric reorganization, whereas regions with stronger dependence on critical subcortical pathways - such as the ventral premotor cortex, the insula (particularly when opercular regions are also involved), and parts of the temporal lobe connected via the arcuate and superior longitudinal fasciculi - show more limited plasticity and frequently retain essential language function within or near the tumor, thereby constraining surgical resection (Nieberlein et al., 2023; Leonard et al., 2024; Herbet and Duffau, 2020). Those findings align with the results of our series in which patients with a combined lesion of the FAT and AF show a low potential for recovery.

Limitations

Study limitations are its retrospective design, including missing data and variability in DTI acquisition. Most patients underwent either pre- or postoperative DTI, limiting cohort consistency. The small sample size, although comparable to other studies (Young et al., 2020; Tuncer et al., 2021; Kinoshita et al., 2015; Herbet and Duffau, 2020), restricts generalizability. Results are exploratory and should be interpreted cautiously.

Another limitation is the use of diffusion MRI tractography to assess white matter tract involvement. Although tractography is widely used for preoperative planning and anatomical visualization, its accuracy may be affected by tumor-related changes such as infiltration, edema, mass effect, tract displacement, and crossing fibers. Consequently, reconstructed fiber tracts represent an approximation of the underlying anatomy rather than a direct measure of white matter integrity. In addition, postoperative tract involvement was inferred by combining preoperative tractography with postoperative structural MRI, which may be affected by brain shift, postoperative edema, cavity collapse, blood products, and other surgery-related distortions, potentially altering the spatial relationship between the resection cavity and the reconstructed fiber tracts. Therefore, the observed associations between tract involvement and postoperative language deficits should be interpreted with appropriate caution. Longitudinal postoperative diffusion MRI acquired after recovery from acute postoperative changes would provide a more reliable assessment of tract integrity but was not routinely available in this retrospective cohort.

A further limitation is that tract involvement was determined using a clinically established deterministic tractography workflow rather than a hypothesis-independent disconnectome approach, which may have limited the detection of alternative disconnection patterns despite reflecting current neurosurgical practice.

Right hemispheric lesions were excluded, although they may contribute to FAT involvement. The focus on left-sided lesions limits generalizability, but bilateral involvement will be explored in future studies. Additionally, we did not analyze other relevant tracts such as the inferior fronto-occipital fasciculus (Giampiccolo and Duffau, 2022).

Further, speech assessments were individualized, limiting standardization and interindividual comparisons. Future studies will address higher neuropsychological functions like processing speed. The absence of standardized postoperative speech-language pathology evaluations limited the ability to retrospectively distinguish between established diagnostic entities, necessitating a descriptive classification of observed speech and language impairments.

Preoperative language deficits were not correlated with FAT integrity due to a small sample. Although preoperative and lasting deficits overlapped, tumor invasion of white matter tracts, particularly the FAT, likely accounts for deficits, though surgical contributions must also be considered. The increased incidence of immediate postoperative deficits suggests surgical disruption of white matter pathways. The higher incidence of immediate postoperative deficits suggests that surgical manipulation of white matter pathways contributes to these impairments, but this interpretation requires confirmation in larger prospective studies.

Tumor heterogeneity, particularly with respect to lesion size, represents an additional limitation of this study. Variability in tumor volume may have influenced both the extent of resection and the degree of white matter involvement, thereby potentially affecting functional outcomes and limiting the comparability of cases within the cohort.