Computer-Aided Segmentation of the Mid-Brain in Trans-Cranial Ultrasound Images

Multiple Kernel Learning Based Classification of Parkinson's Disease With Multi-Modal Transcranial Sonography

Parkinson's Disease (PD) is the most common motor neurodegenerative disease in elderly population. Transcranial sonography (TCS) has become a popular imaging tool for diagnosis of PD in clinical practice. Moreover, several pioneering work have developed the computer-aided diagnosis (CAD) for PD with the transcranial B-mode sonography (TBS). It is worth noting that TCS not only has the TBS modality, but also can image the blood flow of major cerebral arteries, which is named transcranial Doppler sonography (TDS). TDS also has been applied to evaluate PD patients with orthostatic hypotension. However, the TDS-based CAD for PD has not been investigated. Since TBS and TDS provide the complementary structural and functional information about brain, it is feasible to develop a multi-modal TCS-based CAD for PD by combining both TBS and TDS. Therefore, in this work, we propose a multiple kernel learning (MKL) based CAD for PD with multi-modal TCS imaging. Particularly, the statistical and texture features are extracted from the midbrain region from TBS images, and the features about blood flow are calculated from the spectrum curves in TDS. The multi-modal features are then fed to a MKL classifier for classification of PD. The experimental results show that the multi-modal TCS-based method outperforms both the single-modal TBS- and TDS-based algorithm, which suggests the feasibility and effectiveness of combining TBS and TDS for diagnosis of PD.

Transcranioplasty Ultrasound Through a Sonolucent Cranial Implant Made of Polymethyl Methacrylate: Phantom Study Comparing Ultrasound, Computed Tomography, and Magnetic Resonance Imaging

BACKGROUND

Current methods of transcranial diagnostic ultrasound imaging are limited by the skull's acoustic properties. Craniotomy, craniectomy, and cranioplasty procedures present opportunities to circumvent these limitations by substituting autologous bone with synthetic cranial implants composed of sonolucent biomaterials.

OBJECTIVE

This study examined the potential to image the brain using transcranioplasty ultrasound (TCU) through a sonolucent cranial implant.

MATERIALS AND METHODS

A validated adult brain phantom was imaged using computed tomography (CT), magnetic resonance imaging (MRI), and ultrasound without an implant. Next, for experimental comparison, TCU was performed through a sonolucent implant composed of clear polymethyl methacrylate.

RESULTS

All imaging modalities successfully revealed elements of the brain phantom, including the bilateral ventricular system, the falx cerebri, and a deep hyperdense mass representing a brain tumor or hematoma. In addition, ultrasound images were captured which closely resembled axial images obtained with both CT and MRI.

CONCLUSION

The results obtained in this first-ever, preclinical, phantom study suggest TCU is now a viable immediate and long-term diagnostic imaging modality deserving of further clinical investigation.

Multiple Empirical Kernel Mapping Based Broad Learning System for Classification of Parkinson's Disease With Transcranial Sonography

Transcranial sonography (TCS) has become more popular for diagnosis of Parkinson's disease (PD), and the TCS-based computer-aided diagnosis (CAD) for PD also attracts considerable attention, in which classifier is a critical component. Broad learning system (BLS) is a newly proposed single layer feedforward neural network for classification. In BLS, the original input features are mapped to several new feature representations to form the feature nodes, and then these mapped features are expanded to enhancement nodes by random mapping in a wide sense. However, random mapping performed for enhancement nodes is too simple and the generated features lack interpretability together with relative low representation. In this work, we propose a multiple empirical kernel mapping (MEKM) based BLS (MEKM-BLS) algorithm, which adopts MEKM to map the data of feature nodes to enhancement nodes. MEKM-BLS then has more meaningful enhancement layer in feedforward neural network. Moreover, the experiment for PD diagnosis with TCS shows that MEKM-BLS achieves superior performance to the original BLS algorithm.

Transcranial Sonography Based Diagnosis Of Parkinson's Disease Via Cascaded Kernel RVFL

The transcranial sonography (TCS) based computer-aided diagnosis (CAD) for Parkinson's disease (PD) has attracted considerable attention. The learning using privileged information (LUPI) is a new learning paradigm, in which, the privileged information (PI) is only available for model training, but unavailable in the testing stage. The Random vector functional link network plus (RVFL+) algorithm is a newly proposed LUPI algorithm, which has shown its effectiveness for classification task. Moreover, the kernel-based RVFL+ (KRVFL+) has been proposed to overcome the randomness in RVFL+. In this work, we propose a cascaded KRVFL+ (cKRVFL+) algorithm for the single-modal TCS-based PD diagnosis. The predicted value of the former KRVFL+ classifier is adopted as the PI for the current KRVFL+, and only the KRVFL+ in the last layer is finally used as classifiers during the testing stage. This cascaded structure progressively promotes the discrimination performance of KRVFL+ classifier. The experimental results show the effectiveness of the cascaded LUPI classifier framework for single-modality TCS based diagnosis of PD, and the proposed cKRVFL+ algorithm achieves the best performance.

Transcranial Ultrasonographic Image Analysis System for Decision Support in Parkinson Disease

OBJECTIVES

Transcranial ultrasonography (US) is a relatively new neuroimaging modality proposed for early diagnostics of Parkinson disease (PD). The main limitation of transcranial US image-based diagnostics is a high degree of subjectivity caused by low quality of the transcranial images. The article presents a developed image analysis system and evaluates the potential of automated image analysis on transcranial US.

METHODS

The system consists of algorithms for the segmentation and assessment of informative brain regions (midbrain and substantia nigra) and a decision support subsystem, which is equipped with 64 classification algorithms. Transcranial US images of 191 participants (118 patients with a clinical PD diagnosis and 73 healthy control participants) were analyzed.

RESULTS

The diagnostic sensitivity and specificity achieved by the proposed system were 85% and 75%, respectively.

CONCLUSIONS

Digital transcranial US image analysis is challenging, and the application of a such system as the sole instrument for decisions in clinical practice remains inconclusive. However, the proposed system could be used as a supplementary tool for automated assessment of US parameters for decision support in PD diagnostics and to reduce observer variability.