Compensatory Language Reorganization in Two Distinct Stroke Populations

Post by Giulia Baracchini 

What's the science?

Our brain is organized in a number of large-scale networks, so lesions in particular brain areas due to stroke will affect functionally connected regions as well. Post-stroke recovery mechanisms (i.e. brain plasticity) will similarly depend on lesion site and network architecture. How this framework applies to stroke in the context of language impairment (i.e., aphasia), how this framework varies based on lesion site and how it differently impacts acute (1 week after stroke onset), subacute (1-2 weeks) and chronic (> 6 months) post-stroke recovery phases remains unknown. This week in Brain, Stockert and colleagues used functional magnetic resonance imaging (fMRI) to characterize and directly compare language plasticity mechanisms from acute to chronic post-stroke phases in a longitudinal cohort of left temporo-parietal and frontal stroke patients. 

How did they do it?

Thirty-four first ischaemic stroke patients were included in the study, of which 17 showed a left frontal lesion and 17 a left temporo-parietal lesion. Seventeen healthy age-matched controls were also enrolled in the study. All participants underwent fMRI. While in the scanner, participants completed an auditory comprehension task, in which they were asked to actively listen to short German sentences played both in regular order and in reverse. The authors compared activity in language-related brain networks between the frontal lesion stroke patients, the temporo-parietal lesion stroke patients, and controls. Language brain networks were defined as comprising (i) language-relevant areas, (ii) perilesional regions (i.e., 3-15 mm around the lesioned tissue) and (iii) lesion-homolog areas (contralateral areas to the lesioned ones). Importantly, the two patient groups were scanned on three separate occasions, during acute (1 week after stroke onset), subacute (1-2 weeks) and chronic (> 6 months) post-stroke phases. This allowed the authors to investigate language reorganization mechanisms as a function of time post-stroke onset. 

In addition to fMRI, both patient groups completed — for each of the three post-stroke phases — the Aachen Aphasia Test. This test provides a behavioural measure of language production and comprehension performance and quantifies post-stroke language improvement. The authors measured the link between these behavioural scores and brain activation (via fMRI) within and across the two patient groups.

What did they find?

When compared to healthy controls, temporo-parietal patients demonstrated an initial decrease in activity over a vast network including temporo-parietal regions and perilesional frontal areas. This initial global network dysregulation disappeared during the subacute phase and an increase in activity in bilateral frontal areas became apparent, suggesting the restoration of functional connections between temporo-parietal and frontal areas. In contrast, frontal patients in the acute phase reported local dysregulation effects within frontal areas and otherwise preserved perilesional (i.e. left temporo-parietal) and lesion-homolog areas (i.e. right frontal). These findings highlight the importance of considering the lesion site when evaluating early post-stroke mechanisms of language network plasticity. Similarly, when all post-stroke stages were considered together, only frontal patients showed recruitment of lesion-homolog regions, providing further evidence for lesion-dependent reorganization effects.

Lesion-independent reorganization mechanisms were also present during later post-stroke phases. Both temporo-parietal and frontal patients showed increased recruitment of perilesional areas and non-language-specific cognitive control networks (i.e. domain-general networks: cingulo-opercular and fronto-parietal networks) in subacute stages, followed by the recruitment of left language-specific temporal areas (anterior and posterior temporal areas) in chronic phases. These findings suggest that post-stroke functional reorganization occurs in pre-existing functionally connected networks and varies across time after stroke onset. Finally, an overall improvement in behavioural performance for both patient groups was seen, and this improvement was associated with activation in language-specific networks (in both groups) and domain-general networks (in temporo-parietal patients). This finding reflects compensatory mechanisms of language-related and domain-general networks during post-stroke recovery. 

What's the impact?

This study is the first to longitudinally assess post-stroke recovery mechanisms in two distinct stroke populations — temporo-parietal and frontal stroke patients — and directly compare their brain activation patterns. The results speak in favor of a network model for language reorganization that depends on lesion-site and varies based on time post-stroke onset. These novel findings will help to guide post-stroke aphasia rehabilitation strategies, such as non-invasive brain stimulation.

Stockert et al. Dynamics of language reorganization after left temporo-parietal and frontal stroke (2020). Access the original scientific publication here.

Long-Lasting Memory Needs New Myelin Formation

Post by D. Chloe Chung

What's the science?

Our experiences can be recorded in the brain and persist as long-lasting or “remote” memory, through a process called memory consolidation, in which neuronal networks are reorganized over time to structurally represent certain memories. While several neural circuits have been suggested to be potentially responsible for memory consolidation, we still do not fully understand what is happening at the cellular level when particular experiences become solidified as a remote memory. One hint we can find from past studies is that memory functions can be mediated by oligodendrocytes, glial cells that form an insulating fatty tissue called myelin that wraps around axons of the neurons to help efficient neuronal transmission. This week in Nature Neuroscience, Pan and colleagues show that freshly generated oligodendrocytes and their new myelin formation are highly important in consolidating fearful experiences as remote fear memory.

How did they do it?

To make mice in the study learn certain experiences and remember them, the authors used fear conditioning. Fear conditioning trains mice to associate a fearful stimulus like an electric foot shock to neutral cues such as a particular smell or noise. If the mice later freeze (out of fear) after simply being placed in the same chamber even without experiencing a foot shock, this indicates that the mice correctly learned to remember a fearful experience. In this study, right before the fear conditioning, mice were injected with a chemical that incorporates into DNA of dividing cells, so that cells that are newly formed after the injection can be readily identified. 

Brains of these mice were examined either a day or a month after fear learning to evaluate various changes in the brain at different timepoints. To further explore how important new myelin-forming oligodendrocytes are in storing fear memory, the authors used a genetically engineered mouse model that can be induced to stop the production of new oligodendrocytes at any desired timepoint. These transgenic mice underwent the same memory test to determine the effects of new myelin formation on memory consolidation. As another approach, the authors also used a drug called clemastine that can induce new myelin formation and tested how it can change the ability of fear-conditioned mice in recalling remote fear memory.

What did they find?

The authors first observed that mice that successfully learned a fearful experience produced cells with the potential to become oligodendrocytes in the medial prefrontal cortex (mPFC), one of the brain regions important in storing long-term memory. Over the course of several weeks, these cells further matured and became oligodendrocytes that can actively myelinate axons. Interestingly, the authors found that mice that were genetically inhibited to produce new oligodendrocytes, failed to effectively recall a fearful experience from a month ago, as they did not show freezing behaviors like normal mice. Since these transgenic mice were able to remember a fear experience that happened a day before, the authors speculated that new myelin formation is important specifically for the long-term storage of fear memory. Upon closely examining the cells in the genetically inhibited mice that cannot form new myelin, the authors found that neurons in the memory consolidation brain regions were not efficiently activated, as shown by greatly reduced expressions of immediate early genes that are normally activated with external stimuli. Mice injected with myelination-promoting clemastine showed remote fear memory that was significantly more persistent and stabilized than the control group. This observation, accompanied by increased expression of immediate early genes, suggests that new myelin formation can substantially preserve remote fear memory.

What’s the impact?

This study presents the first evidence that effective storage of fear memories over a long period of time requires newly generated oligodendrocytes that can provide fresh myelin for neurons. This study is particularly interesting as it showed that myelination, which has been thought to mostly occur during the early development of the brain, is, in fact, a very dynamic process that can critically help to preserve long-lasting memories in the adult brain. Findings from this study may ultimately provide useful insights in developing effective treatment options for neurological diseases like post-traumatic stress disorder (PTSD), a disorder associated with the abnormal, intense recall of fear memories.

Pan et al. Preservation of a remote fear memory requires new myelin formation. Nature Neuroscience (2020). Access the original scientific publication here.

A New Clinical Syndrome: Trigeminal Neuralgia Associated with a Brain Lesion

Post by Kasey Hemington

What's the science?

Trigeminal Neuralgia (TN) is an excruciatingly painful condition. Patients experience severe pain attacks in areas associated with the trigeminal nerve, which innervates the face. In some cases, TN is associated with demyelination (damage to myelin – a protective covering for the nerve) either due to compression of the trigeminal nerve itself or secondary to Multiple Sclerosis (MS-TN), which can cause demyelinating plaques in the brainstem. A few case studies have noted that some TN patients have lesions of the brainstem, but do not have Multiple Sclerosis. However, this unique patient sub-group has not been extensively studied. This week in PAIN, Tohyama and colleagues used magnetic resonance imaging (MRI) and clinical evidence to define a new syndrome: TN associated with solitary pontine lesion (SPL-TN; ‘pontine’ refers to the pons, a region of the brainstem)

How did they do it?

The authors examined clinical records and MRI brain scans for 481 TN patients who underwent neurosurgical TN treatment (commonly gamma knife surgery, a non-invasive type of surgery that uses beams of radiation). All patients underwent an anatomical MRI (T1-weighted), and a subset of patients also underwent a diffusion-weighted MRI. SPL-TN was defined based on an idiopathic TN diagnosis (e.g. TN not secondary to Multiple Sclerosis), a single lesion along the trigeminal nerve pathway, and no other brain lesions. To characterize the lesions, the authors used an anatomical MRI and mapped the lesion by hand in each patient before comparing lesion distribution and area across subjects. The authors hypothesized SPL-TN patients to be surgical treatment non-responders, and defined treatment non-response as having undergone three or more surgical procedures, or having undergone one surgical procedure without experiencing substantial pain relief. Lack of relief was defined as <75% pain reduction using an 11-point pain Numerical Rating Scale and a score of 4 or higher on the Barrow Neurological Institute Scale, which measures the frequency of medication use for pain control.

From diffusion-weighted MRI scans of the lesions, the authors calculated metrics including fractional anisotropy, mean diffusivity, axial diffusivity, and radial diffusivity, which reflect the brain’s white matter microstructure, neuroinflammation, myelination, and axonal integrity, respectively. To act as a healthy control group, the authors recruited healthy individuals, age- and sex-matched to SPL-TN patients for whom anatomical and diffusion-weighted MRI brain scans were available. The healthy control group underwent these same brain scans.

What did they find?

Upon reviewing the 481 TN patient records, the authors found 24 cases of SPL-TN. Eighteen of those patients had clinical follow-up information available post-surgery and it was determined that 17/18 of those patients did not respond to surgical treatment, suggesting non-response to surgical treatment is characteristic of SPL-TN. For these 17 patients (6 men, 11 women), lesions were mapped and found to be along the trigeminal pathway in the pons on the affected side of the brain in all cases. Overlap amongst patients was maximal in the trigeminal brainstem sensory nuclear complex. Diffusion-weighted MRI scans were available for 11 of the 17 patients, and were used to compare the lesions on the affected side of the brain in SPL-TN patients versus a) the same region on the non-affected side and b) the brains of 11 healthy controls. No differences in fractional anisotropy, mean diffusivity, axial diffusivity, or radial diffusivity were found between the non-affected side in patients and healthy controls, while lower fractional anisotropy and higher mean diffusivity and radial diffusivity were found on the affected side versus the non-affected sides and healthy controls. Additionally, axial diffusivity was higher on the affected side versus the non-affected side. These findings indicate abnormal white matter properties in SPL-TN lesions.

FA = Fractional Anisotropy, MD = Mean Diffusivity, RD = Radial Diffusivity, AD = Axial Diffusivity

SPL-TN patient lesions were also compared to lesions in a group of 17 MS-TN patients, age-matched for the age of TN onset. In MS-TN patients, between 1-5 lesions could be found along the trigeminal nerve pathway. No differences between groups were found in diffusivity metrics when comparing whole lesions. However, when the analysis was restricted precisely to the trigeminal brainstem tract within a lesion, lower fractional anisotropy, and higher axial diffusivity, mean diffusivity, and radial diffusivity were found in SPL-TN lesions versus MS-TN lesions.

What's the impact?

This study is the first to define SPL-TN, a subtype of TN characterized by non-response to surgical treatment and a single brainstem lesion commonly in the trigeminal nucleus. These lesions have changes in diffusivity metrics that characterize abnormal white matter microstructure. The identification of SPL-TN will guide specialized treatment plans for these patients.

Tohyama et al. Trigeminal neuralgia associated with a solitary pontine lesion. PAIN (2020). Access the original scientific publication here.