Registration of Brain CT Images to an MRI Template for the Purpose of Lesion-Symptom Mapping. In: Shen L, Liu T, Yap P, Huang H, Shen D, Westin C, editors. Multimodal Brain Image Analysis. Cham: Springer International Publishing; 2013. pp. 119-28. [DOI: 10.1007/978-3-319-02126-3_12]

A brain CT-based approach for predicting and analyzing stroke-associated pneumonia from intracerebral hemorrhage

Introduction

Stroke-associated pneumonia (SAP) is a common complication of stroke that can increase the mortality rate of patients and the burden on their families. In contrast to prior clinical scoring models that rely on baseline data, we propose constructing models based on brain CT scans due to their accessibility and clinical universality.

Methods

Our study aims to explore the mechanism behind the distribution and lesion areas of intracerebral hemorrhage (ICH) in relation to pneumonia, we utilized an MRI atlas that could present brain structures and a registration method in our program to extract features that may represent this relationship. We developed three machine learning models to predict the occurrence of SAP using these features. Ten-fold cross-validation was applied to evaluate the performance of models. Additionally, we constructed a probability map through statistical analysis that could display which brain regions are more frequently impacted by hematoma in patients with SAP based on four types of pneumonia.

Results

Our study included a cohort of 244 patients, and we extracted 35 features that captured the invasion of ICH to different brain regions for model development. We evaluated the performance of three machine learning models, namely, logistic regression, support vector machine, and random forest, in predicting SAP, and the AUCs for these models ranged from 0.77 to 0.82. The probability map revealed that the distribution of ICH varied between the left and right brain hemispheres in patients with moderate and severe SAP, and we identified several brain structures, including the left-choroid-plexus, right-choroid-plexus, right-hippocampus, and left-hippocampus, that were more closely related to SAP based on feature selection. Additionally, we observed that some statistical indicators of ICH volume, such as mean and maximum values, were proportional to the severity of SAP.

Discussion

Our findings suggest that our method is effective in classifying the development of pneumonia based on brain CT scans. Furthermore, we identified distinct characteristics, such as volume and distribution, of ICH in four different types of SAP.

Improving Patch-Based Convolutional Neural Networks for MRI Brain Tumor Segmentation by Leveraging Location Information

The manual brain tumor annotation process is time consuming and resource consuming, therefore, an automated and accurate brain tumor segmentation tool is greatly in demand. In this paper, we introduce a novel method to integrate location information with the state-of-the-art patch-based neural networks for brain tumor segmentation. This is motivated by the observation that lesions are not uniformly distributed across different brain parcellation regions and that a locality-sensitive segmentation is likely to obtain better segmentation accuracy. Toward this, we use an existing brain parcellation atlas in the Montreal Neurological Institute (MNI) space and map this atlas to the individual subject data. This mapped atlas in the subject data space is integrated with structural Magnetic Resonance (MR) imaging data, and patch-based neural networks, including 3D U-Net and DeepMedic, are trained to classify the different brain lesions. Multiple state-of-the-art neural networks are trained and integrated with XGBoost fusion in the proposed two-level ensemble method. The first level reduces the uncertainty of the same type of models with different seed initializations, and the second level leverages the advantages of different types of neural network models. The proposed location information fusion method improves the segmentation performance of state-of-the-art networks including 3D U-Net and DeepMedic. Our proposed ensemble also achieves better segmentation performance compared to the state-of-the-art networks in BraTS 2017 and rivals state-of-the-art networks in BraTS 2018. Detailed results are provided on the public multimodal brain tumor segmentation (BraTS) benchmarks.

Strategic infarct location for post-stroke cognitive impairment: A multivariate lesion-symptom mapping study

Lesion location is an important determinant for post-stroke cognitive impairment. Although several 'strategic' brain regions have previously been identified, a comprehensive map of strategic brain regions for post-stroke cognitive impairment is lacking due to limitations in sample size and methodology. We aimed to determine strategic brain regions for post-stroke cognitive impairment by applying multivariate lesion-symptom mapping in a large cohort of 410 acute ischemic stroke patients. Montreal Cognitive Assessment at three to six months after stroke was used to assess global cognitive functioning and cognitive domains (memory, language, attention, executive and visuospatial function). The relation between infarct location and cognition was assessed in multivariate analyses at the voxel-level and the level of regions of interest using support vector regression. These two assumption-free analyses consistently identified the left angular gyrus, left basal ganglia structures and the white matter around the left basal ganglia as strategic structures for global cognitive impairment after stroke. A strategic network involving several overlapping and domain-specific cortical and subcortical structures was identified for each of the cognitive domains. Future studies should aim to develop even more comprehensive infarct location-based models for post-stroke cognitive impairment through multicenter studies including thousands of patients.

Peripersonal and extrapersonal visuospatial neglect in different frames of reference: A brain lesion-symptom mapping study

INTRODUCTION

Visuospatial neglect can occur in peripersonal and extrapersonal space. The dorsal visual pathway is hypothesized to be associated with peripersonal, and the ventral pathway with extrapersonal neglect. We aimed to evaluate neural substrates of peripersonal versus extrapersonal neglect, separately for egocentric and allocentric frames of reference.

METHODS

This was a retrospective study, including stroke patients admitted for inpatient rehabilitation. Approximately 1 month post-stroke onset, computerized cancellation (egocentric) and bisection tasks (egocentric and allocentric) were administered at 30 cm and 120 cm. We collected CT or MRI scans and performed voxel-based lesion-symptom mapping for the cancellation, and subtraction analyses for the line bisection task.

RESULTS

We included 98 patients for the cancellation and 129 for the bisection analyses. The right parahippocampal gyrus, hippocampus, and thalamus were associated with egocentric peripersonal neglect as measured with cancellation. These areas were also associated with extrapersonal neglect, together with the right superior parietal lobule, angular gyrus, supramarginal gyrus, lateral occipital cortex, planum temporale and superior temporal gyrus. Lesions in the right parietal, temporal and frontal areas were associated with both peripersonal and extrapersonal egocentric neglect as measured with bisection. For allocentric neglect no clear pattern of associated brain regions was observed.

DISCUSSION

We found right hemispheric anatomical correlates for peripersonal and extrapersonal neglect. However, no brain areas were uniquely associated with peripersonal neglect, meaning we could not conclusively verify the ventral/dorsal hypothesis. Several areas were uniquely associated with egocentric extrapersonal neglect, suggesting that these brain areas can be specifically involved in extrapersonal, but not in peripersonal, attention processes.

Shared and distinct anatomical correlates of semantic and phonemic fluency revealed by lesion-symptom mapping in patients with ischemic stroke

Semantic and phonemic fluency tasks are frequently used to test executive functioning, speed and attention, and access to the mental lexicon. In semantic fluency tasks, subjects are required to generate words belonging to a category (e.g., animals) within a limited time window, whereas in phonemic fluency tasks subjects have to generate words starting with a given letter. Anatomical correlates of semantic and phonemic fluency are currently assumed to overlap in left frontal structures, reflecting shared executive processes, and to be distinct in left temporal and right frontal structures, reflecting involvement of distinct memory processes and search strategies. Definite evidence for this assumption is lacking. To further establish the anatomical correlates of semantic and phonemic fluency, we applied assumption-free voxel-based and region-of-interest-based lesion-symptom mapping in 93 patients with ischemic stroke. Fluency was assessed by asking patients to name animals (semantic), and words starting with the letter N and A (phonemic). Our findings indicate that anatomical correlates of semantic and phonemic fluency overlap in the left inferior frontal gyrus and insula, reflecting shared underlying cognitive processes. Phonemic fluency additionally draws on the left rolandic operculum, which might reflect a search through phonological memory, and the middle frontal gyrus. Semantic fluency additionally draws on left medial temporal regions, probably reflecting a search through semantic memory, and the right inferior frontal gyrus, which might reflect the application of a visuospatial mental imagery strategy in semantic fluency. These findings establish shared and distinct anatomical correlates of semantic and phonemic fluency.

The anatomy of visuospatial construction revealed by lesion-symptom mapping

Visuospatial construction is a complex cognitive operation that is composed of a purely constructional component (visuoconstruction proper), and visuoperceptive, attentional, and decision-making components. The anatomical correlates of visuospatial construction and its cognitive subcomponents are poorly understood. The purpose of the present study was to determine the anatomical correlates of visuospatial construction by applying lesion-symptom mapping in a cohort of 111 patients with first-ever ischemic stroke. We employed the Rey-Osterrieth Complex Figure (ROCF) copy test and the Judgment of Line Orientation (JLO); both tests measure visuoperception, while only the ROCF has a constructional component. We first performed assumption-free voxel-based lesion-symptom mapping, which revealed large shared right hemispheric correlates for the ROCF and JLO in the frontal lobe, superior temporal lobe, and supramarginal gyrus. These shared anatomical correlates reflect the visuoperceptive component of the ROCF and JLO. Anatomical correlates were discordant in the right superior parietal lobule, and angular and middle occipital gyri: lesions in these regions were associated with poor performance on the ROCF, but not the JLO. Secondly, these findings were reproduced with a region of interest-based analysis that yielded a statistically significant correlation between infarct volume in the right inferior and superior parietal, angular and middle occipital cortices, and poor performance on the ROCF, but not the JLO. This discordance in anatomical correlates of the ROCF and JLO reflects the visuoconstructive component of the ROCF. These findings provide new insights in the anatomical correlates of the visuoperceptive and visuoconstructive components of the ROCF and provide evidence for a crucial role of the right inferior and superior parietal, angular and middle occipital gyri in visuoconstruction proper.