Classic models of language processing posit a dorsal stream involving Wernicke's area (posterior temporal cortex) for comprehension and Broca's area (posterior inferior frontal gyrus) for production. However, recent research suggests a crucial role for the anterior lateral temporal lobe (ATL) in language, particularly in semantic processing. The exact contribution of the left ATL remains debated, with some studies finding no semantic deficits following unilateral left ATL lesions, while others report lexical retrieval deficits or semantic errors. Anatomical studies reveal the anterior ATL's connection to the frontal cortex via the temporo-frontal extreme capsule fasciculus (TFexcF), specifically targeting the pars triangularis (area 45) of Broca's region. This TFexcF, a direct monosynaptic pathway, distinguishes itself from the uncinate fasciculus and the multi-component inferior fronto-occipital fasciculus (IFOF). The TFexcF forms a key part of the ventral language network, while the pars opercularis (area 44) connects primarily with the dorsal system. The dorsal system is linked to phonological processing, whereas the ventral system is associated with semantic processing. Electrical stimulation studies show that stimulating area 44 results in speech arrest, whereas area 45 stimulation causes mild disturbances like word-finding difficulties. This study explores a unique case of selective damage to area 45, leaving area 44 intact, to further understand the distinct role of the ventral language system.

The literature on the anterior temporal lobe's (ATL) role in language processing is mixed. Some studies show no significant semantic deficits after unilateral left ATL lesions, while others report lexical-retrieval deficits or semantic error production following lesions to this area, particularly the middle temporal gyrus (MTG). The studies focusing on unilateral lesions in the acute phase, have shown contradictory results. The studies using larger samples of patients in the chronic phase, demonstrated a significant relationship between anterior MTG lesions and semantic error production, while other studies focused on severe word comprehension impairment in relation with left ATL atrophy. Anatomical studies show the existence of a direct monosynaptic pathway connecting the anterior lateral temporal cortex and the frontal cortex, particularly the pars triangularis (area 45) of Broca's area, via the temporo-frontal extreme capsule fasciculus (TFexcF). This contrasts with the multi-component inferior fronto-occipital fasciculus (IFOF), which connects occipital cortex to posterior inferior temporal cortex and then to frontal cortex, which frequently causes confusion in the literature.

This study examined three right-handed male patients with chronic left-hemisphere middle cerebral artery cerebrovascular accidents (CVAs). Patient AA had a lesion restricted to the ventral anterior temporal network, including the anterior superior and middle temporal gyri and pars triangularis (area 45) of the inferior frontal gyrus (IFG), and damage to the TFexcF. Patients MM and TA had lesions affecting the classic dorsal posterior temporo-parietal language system. Diffusion tensor imaging (DTI) was used to reconstruct white matter tracts in all three patients and a control group of 10 healthy males. Neuropsychological assessments included the Boston Diagnostic Aphasia Examination (BDAE) (for oral expression, comprehension, repetition, and reading), Boston Naming Test (BNT), Peabody Picture Vocabulary Test-Revised (PPVT-R), Controlled Oral Word Fluency (COWF) test, Trail Making Test, Pyramids and Palm Trees test (PPT), Taylor Complex Figure test, and various memory tasks. Fractional anisotropy (FA) values were compared across groups using Wilcoxon signed-rank tests.

Patient AA, with the ventral anterior temporal lesion, exhibited fluent and error-free speech but significantly impaired performance on the PPVT-R (receptive vocabulary), semantic fluency, and phonemic fluency tests. These deficits point to difficulties in semantic retrieval. In contrast, patients MM and TA (with dorsal lesions) showed the typical aphasic profile of impaired speech, sentence comprehension, and repetition. The DTI analysis showed reduced FA in the TFexcF for patient AA, and reduced FA in SLF III and AF for patients MM and TA. This differential performance underscores the specific role of the ventral system (involving area 45) in selective controlled retrieval of lexico-semantic information and highlights the dorsal system's involvement in phonological processing, sentence comprehension, and auditory working memory. Specifically, patient AA's preserved performance on the BDAE word comprehension subtest (compared to the PPVT-R) suggests intact basic word comprehension, but difficulties in accessing complex semantic information and retrieval processes. Patients MM and TA showed the typical pattern of fluent aphasia, with impaired auditory comprehension and sentence repetition, suggesting the impairment of phonological and/or somatosensory processing. Patient AA’s unimpaired sentence repetition can be associated to his intact dorsal posterior temporal language system. The preserved auditory working memory in AA further supports the interpretation that the impairment in this patient is limited to the ability of retrieving semantic representation. The phonemic errors and neologisms in MM and TA supports the important role of the dorsal network (area 44, AF, and SLF III) in the speech production process, particularly the phonological aspects of speech output.

The findings strongly support the dual-stream model of language, differentiating the contributions of the ventral and dorsal pathways. The selective impairment in controlled semantic retrieval observed in patient AA, whose lesion spared area 44, implicates area 45 in this higher-order cognitive function. This challenges previous interpretations of area 45's role based on electrical stimulation studies. The contrast between patient AA's profile and that of patients MM and TA confirms the specific role of the ventral language system in lexico-semantic processing, and supports the importance of the dorsal language system for phonological processing and auditory working memory. The observed dissociation between single word comprehension (BDAE) and complex word comprehension (PPVT-R) in patient AA further supports this distinction. The current findings are congruent with those obtained from studies of patients with primary progressive aphasia (PPA) where dissociation between single word comprehension and sentence comprehension were observed.

This study provides compelling evidence for the distinct roles of the ventral and dorsal language networks. Damage to the ventral system, specifically area 45 and its connections via the TFexcF, selectively impairs controlled semantic retrieval, highlighting its importance for lexico-semantic processing and language comprehension. Conversely, the dorsal system plays a crucial role in phonological processing, sentence comprehension, and auditory working memory. Future research should investigate the specific cognitive mechanisms underlying area 45's function and explore the dynamic interactions between the ventral and dorsal streams.

The study's limitations include the small sample size (three patients) and the use of a relatively low number of directions and b-values in the DTI acquisition. Furthermore, the study focuses on the chronic phase of stroke recovery, limiting the insights into potential functional plasticity. Improved distortion and correction algorithms in DTI analysis could enhance the accuracy of tract reconstructions.