Global brain network modularity dynamics after local optic nerve damage following noninvasive brain stimulation: an EEG-tracking study.

Tightly connected clusters of nodes, called communities, interact in a time-dependent manner in brain functional connectivity networks (FCN) to support complex cognitive functions. However, little is known if and how different nodes synchronize their neural interactions to form functional communities ("modules") during visual processing and if and how this modularity changes postlesion (progression or recovery) following neuromodulation. Using the damaged optic nerve as a paradigm, we now studied brain FCN modularity dynamics to better understand module interactions and dynamic reconfigurations before and after neuromodulation with noninvasive repetitive transorbital alternating current stimulation (rtACS). We found that in both patients and controls, local intermodule interactions correlated with visual performance. However, patients' recovery of vision after treatment with rtACS was associated with improved interaction strength of pathways linked to the attention module, and it improved global modularity and increased the stability of FCN. Our results show that temporal coordination of multiple cortical modules and intermodule interaction are functionally relevant for visual processing. This modularity can be neuromodulated with tACS, which induces a more optimal balanced and stable multilayer modular structure for visual processing by enhancing the interaction of neural pathways with the attention network module.

Deformable 3D/3D CT-to-digital-tomosynthesis image registration in image-guided bronchoscopy interventions.

Traditional navigational bronchoscopy procedures rely on preprocedural computed tomography (CT) and intraoperative chest radiography and cone-beam CT (CBCT) to biopsy peripheral lung lesions. This navigational approach is challenging due to the projective nature of radiography, and the high radiation dose, long imaging time, and large footprints of CBCT. Digital tomosynthesis (DTS) is considered an attractive alternative combining the advantages of radiography and CBCT. Only the depth resolution cannot match a full CBCT image due to the limited angle acquisition. To address this issue, preoperative CT is a good auxiliary in guiding bronchoscopy interventions. Nevertheless, CT-to-body divergence caused by anatomic changes and respiratory motion, hinders the effective use of CT imaging. To mitigate CT-to-body divergence, we propose a novel deformable 3D/3D CT-to-DTS registration algorithm employing a multistage, multiresolution approach and using affine and elastic B-spline transformation models with bone and lung mask images. A multiresolution strategy with a Gaussian image pyramid and a multigrid strategy within the B-spline model are applied. The normalized correlation coefficient is included in the cost function for the affine model and a multimetric weighted cost function is used for the B-spline model, with weights determined heuristically. Tested on simulated and real patient bronchoscopy data, the algorithm yields promising results. Assessed qualitatively by visual inspection and quantitatively by computing the Dice coefficient (DC) and the average symmetric surface distance (ASSD), the algorithm achieves mean DC of 0.82±0.05 and 0.74±0.05, and mean ASSD of 0.65±0.29mm and 0.93±0.43mm for simulated and real data, respectively. This algorithm lays the groundwork for CT-aided intraoperative DTS imaging in image-guided bronchoscopy interventions with future studies focusing on automated metric weight setting.

Exploration of Interpretability Techniques for Deep COVID-19 Classification Using Chest X-ray Images.

The outbreak of COVID-19 has shocked the entire world with its fairly rapid spread, and has challenged different sectors. One of the most effective ways to limit its spread is the early and accurate diagnosing of infected patients. Medical imaging, such as X-ray and computed tomography (CT), combined with the potential of artificial intelligence (AI), plays an essential role in supporting medical personnel in the diagnosis process. Thus, in this article, five different deep learning models (ResNet18, ResNet34, InceptionV3, InceptionResNetV2, and DenseNet161) and their ensemble, using majority voting, have been used to classify COVID-19, pneumoniæ and healthy subjects using chest X-ray images. Multilabel classification was performed to predict multiple pathologies for each patient, if present. Firstly, the interpretability of each of the networks was thoroughly studied using local interpretability methods-occlusion, saliency, input X gradient, guided backpropagation, integrated gradients, and DeepLIFT-and using a global technique-neuron activation profiles. The mean micro F1 score of the models for COVID-19 classifications ranged from 0.66 to 0.875, and was 0.89 for the ensemble of the network models. The qualitative results showed that the ResNets were the most interpretable models. This research demonstrates the importance of using interpretability methods to compare different models before making a decision regarding the best performing model.

Sinogram upsampling using Primal-Dual UNet for undersampled CT and radial MRI reconstruction.

Computed tomography (CT) and magnetic resonance imaging (MRI) are two widely used clinical imaging modalities for non-invasive diagnosis. However, both of these modalities come with certain problems. CT uses harmful ionising radiation, and MRI suffers from slow acquisition speed. Both problems can be tackled by undersampling, such as sparse sampling. However, such undersampled data leads to lower resolution and introduces artefacts. Several techniques, including deep learning based methods, have been proposed to reconstruct such data. However, the undersampled reconstruction problem for these two modalities was always considered as two different problems and tackled separately by different research works. This paper proposes a unified solution for both sparse CT and undersampled radial MRI reconstruction, achieved by applying Fourier transform-based pre-processing on the radial MRI and then finally reconstructing both modalities using sinogram upsampling combined with filtered back-projection. The Primal-Dual network is a deep learning based method for reconstructing sparsely-sampled CT data. This paper introduces Primal-Dual UNet, which improves the Primal-Dual network in terms of accuracy and reconstruction speed. The proposed method resulted in an average SSIM of 0.932±0.021 while performing sparse CT reconstruction for fan-beam geometry with a sparsity level of 16, achieving a statistically significant improvement over the previous model, which resulted in 0.919±0.016. Furthermore, the proposed model resulted in 0.903±0.019 and 0.957±0.023 average SSIM while reconstructing undersampled brain and abdominal MRI data with an acceleration factor of 16, respectively - statistically significant improvements over the original model, which resulted in 0.867±0.025 and 0.949±0.025. Finally, this paper shows that the proposed network not only improves the overall image quality, but also improves the image quality for the regions-of-interest: liver, kidneys, and spleen; as well as generalises better than the baselines in presence the of a needle.

MICDIR: Multi-scale inverse-consistent deformable image registration using UNetMSS with self-constructing graph latent.

Image registration is the process of bringing different images into a common coordinate system - a technique widely used in various applications of computer vision, such as remote sensing, image retrieval, and, most commonly, medical imaging. Deep learning based techniques have been applied successfully to tackle various complex medical image processing problems, including medical image registration. Over the years, several image registration techniques have been proposed using deep learning. Deformable image registration techniques such as Voxelmorph have been successful in capturing finer changes and providing smoother deformations. However, Voxelmorph, as well as ICNet and FIRE, do not explicitly encode global dependencies (i.e. the overall anatomical view of the supplied image) and, therefore, cannot track large deformations. In order to tackle the aforementioned problems, this paper extends the Voxelmorph approach in three different ways. To improve the performance in case of small as well as large deformations, supervision of the model at different resolutions has been integrated using a multi-scale UNet. To support the network to learn and encode the minute structural co-relations of the given image-pairs, a self-constructing graph network (SCGNet) has been used as the latent of the multi-scale UNet - which can improve the learning process of the model and help the model to generalise better. And finally, to make the deformations inverse-consistent, cycle consistency loss has been employed. On the task of registration of brain MRIs, the proposed method achieved significant improvements over ANTs and VoxelMorph, obtaining a Dice score of 0.8013 ± 0.0243 for intramodal and 0.6211 ± 0.0309 for intermodal, while VoxelMorph achieved 0.7747 ± 0.0260 and 0.6071 ± 0.0510, respectively.

Liver segmentation using Turbolift learning for CT and cone-beam C-arm perfusion imaging.

Model-based reconstruction employing the time separation technique (TST) was found to improve dynamic perfusion imaging of the liver using C-arm cone-beam computed tomography (CBCT). To apply TST using prior knowledge extracted from CT perfusion data, the liver should be accurately segmented from the CT scans. Reconstructions of primary and model-based CBCT data need to be segmented for proper visualisation and interpretation of perfusion maps. This research proposes Turbolift learning, which trains a modified version of the multi-scale Attention UNet on different liver segmentation tasks serially, following the order of the trainings CT, CBCT, CBCT TST - making the previous trainings act as pre-training stages for the subsequent ones - addressing the problem of limited number of datasets for training. For the final task of liver segmentation from CBCT TST, the proposed method achieved an overall Dice scores of 0.874±0.031 and 0.905±0.007 in 6-fold and 4-fold cross-validation experiments, respectively - securing statistically significant improvements over the model, which was trained only for that task. Experiments revealed that Turbolift not only improves the overall performance of the model but also makes it robust against artefacts originating from the embolisation materials and truncation artefacts. Additionally, in-depth analyses confirmed the order of the segmentation tasks. This paper shows the potential of segmenting the liver from CT, CBCT, and CBCT TST, learning from the available limited training data, which can possibly be used in the future for the visualisation and evaluation of the perfusion maps for the treatment evaluation of liver diseases.

Classification of brain tumours in MR images using deep spatiospatial models.

A brain tumour is a mass or cluster of abnormal cells in the brain, which has the possibility of becoming life-threatening because of its ability to invade neighbouring tissues and also form metastases. An accurate diagnosis is essential for successful treatment planning, and magnetic resonance imaging is the principal imaging modality for diagnosing brain tumours and their extent. Deep Learning methods in computer vision applications have shown significant improvement in recent years, most of which can be credited to the fact that a sizeable amount of data is available to train models, and the improvements in the model architectures yield better approximations in a supervised setting. Classifying tumours using such deep learning methods has made significant progress with the availability of open datasets with reliable annotations. Typically those methods are either 3D models, which use 3D volumetric MRIs or even 2D models considering each slice separately. However, by treating one spatial dimension separately or by considering the slices as a sequence of images over time, spatiotemporal models can be employed as "spatiospatial" models for this task. These models have the capabilities of learning specific spatial and temporal relationships while reducing computational costs. This paper uses two spatiotemporal models, ResNet (2+1)D and ResNet Mixed Convolution, to classify different types of brain tumours. It was observed that both these models performed superior to the pure 3D convolutional model, ResNet18. Furthermore, it was also observed that pre-training the models on a different, even unrelated dataset before training them for the task of tumour classification improves the performance. Finally, Pre-trained ResNet Mixed Convolution was observed to be the best model in these experiments, achieving a macro F1-score of 0.9345 and a test accuracy of 96.98%, while at the same time being the model with the least computational cost.

ReconResNet: Regularised residual learning for MR image reconstruction of Undersampled Cartesian and Radial data.

MRI is an inherently slow process, which leads to long scan time for high-resolution imaging. The speed of acquisition can be increased by ignoring parts of the data (undersampling). Consequently, this leads to the degradation of image quality, such as loss of resolution or introduction of image artefacts. This work aims to reconstruct highly undersampled Cartesian or radial MR acquisitions, with better resolution and with less to no artefact compared to conventional techniques like compressed sensing. In recent times, deep learning has emerged as a very important area of research and has shown immense potential in solving inverse problems, e.g. MR image reconstruction. In this paper, a deep learning based MR image reconstruction framework is proposed, which includes a modified regularised version of ResNet as the network backbone to remove artefacts from the undersampled image, followed by data consistency steps that fusions the network output with the data already available from undersampled k-space in order to further improve reconstruction quality. The performance of this framework for various undersampling patterns has also been tested, and it has been observed that the framework is robust to deal with various sampling patterns, even when mixed together while training, and results in very high quality reconstruction, in terms of high SSIM (highest being 0.990 ± 0.006 for acceleration factor of 3.5), while being compared with the fully sampled reconstruction. It has been shown that the proposed framework can successfully reconstruct even for an acceleration factor of 20 for Cartesian (0.968 ± 0.005) and 17 for radially (0.962 ± 0.012) sampled data. Furthermore, it has been shown that the framework preserves brain pathology during reconstruction while being trained on healthy subjects.

StRegA: Unsupervised anomaly detection in brain MRIs using a compact context-encoding variational autoencoder.

Expert interpretation of anatomical images of the human brain is the central part of neuroradiology. Several machine learning-based techniques have been proposed to assist in the analysis process. However, the ML models typically need to be trained to perform a specific task, e.g., brain tumour segmentation or classification. Not only do the corresponding training data require laborious manual annotations, but a wide variety of abnormalities can be present in a human brain MRI - even more than one simultaneously, which renders a representation of all possible anomalies very challenging. Hence, a possible solution is an unsupervised anomaly detection (UAD) system that can learn a data distribution from an unlabelled dataset of healthy subjects and then be applied to detect out-of-distribution samples. Such a technique can then be used to detect anomalies - lesions or abnormalities, for example, brain tumours, without explicitly training the model for that specific pathology. Several Variational Autoencoder (VAE) based techniques have been proposed in the past for this task. Even though they perform very well on controlled artificially simulated anomalies, many of them perform poorly while detecting anomalies in clinical data. This research proposes a compact version of the "context-encoding" VAE (ceVAE) model, combined with pre and post-processing steps, creating a UAD pipeline (StRegA), which is more robust on clinical data and shows its applicability in detecting anomalies such as tumours in brain MRIs. The proposed pipeline achieved a Dice score of 0.642 ± 0.101 while detecting tumours in T2w images of the BraTS dataset and 0.859 ± 0.112 while detecting artificially induced anomalies, while the best performing baseline achieved 0.522 ± 0.135 and 0.783 ± 0.111, respectively.

Adaptive and Maladaptive Brain Functional Network Reorganization After Stroke in Hemianopia Patients: An Electroencephalogram-Tracking Study.

Objective: Hemianopia after occipital stroke is believed to be mainly due to local damage at or near the lesion site. However, magnetic resonance imaging studies suggest functional connectivity network (FCN) reorganization also in distant brain regions. Because it is unclear whether reorganization is adaptive or maladaptive, compensating for, or aggravating vision loss, we characterized FCNs electrophysiologically to explore local and global brain plasticity and correlated FCN reorganization with visual performance. Methods: Resting-state electroencephalography (EEG) was recorded in chronic, unilateral stroke patients and healthy age-matched controls (n = 24 each). This study was approved by the local ethics committee. The correlation of oscillating EEG activity was calculated with the imaginary part of coherence between pairs of regions of interest, and FCN graph theory metrics (degree, strength, clustering coefficient) were correlated with stimulus detection and reaction time. Results: Stroke brains showed altered FCNs in the alpha- and low beta-band in numerous occipital, temporal brain structures. On a global level, FCN had a less efficient network organization whereas on the local level node networks were reorganized especially in the intact hemisphere. Here, the occipital network was 58% more rigid (with a more "regular" network structure) whereas the temporal network was 32% more efficient (showing greater "small-worldness"), both of which correlated with worse or better visual processing, respectively. Conclusions: Occipital stroke is associated with both local and global FCN reorganization, but this can be both adaptive and maladaptive. We propose that the more "regular" FCN structure in the intact visual cortex indicates maladaptive plasticity, where less processing efficacy with reduced signal/noise ratio may cause the perceptual deficits in the intact visual field (VF). In contrast, reorganization in intact temporal brain regions is presumably adaptive, possibly supporting enhanced peripheral movement perception.

DS6, Deformation-Aware Semi-Supervised Learning: Application to Small Vessel Segmentation with Noisy Training Data.

Blood vessels of the brain provide the human brain with the required nutrients and oxygen. As a vulnerable part of the cerebral blood supply, pathology of small vessels can cause serious problems such as Cerebral Small Vessel Diseases (CSVD). It has also been shown that CSVD is related to neurodegeneration, such as Alzheimer's disease. With the advancement of 7 Tesla MRI systems, higher spatial image resolution can be achieved, enabling the depiction of very small vessels in the brain. Non-Deep Learning-based approaches for vessel segmentation, e.g., Frangi's vessel enhancement with subsequent thresholding, are capable of segmenting medium to large vessels but often fail to segment small vessels. The sensitivity of these methods to small vessels can be increased by extensive parameter tuning or by manual corrections, albeit making them time-consuming, laborious, and not feasible for larger datasets. This paper proposes a deep learning architecture to automatically segment small vessels in 7 Tesla 3D Time-of-Flight (ToF) Magnetic Resonance Angiography (MRA) data. The algorithm was trained and evaluated on a small imperfect semi-automatically segmented dataset of only 11 subjects; using six for training, two for validation, and three for testing. The deep learning model based on U-Net Multi-Scale Supervision was trained using the training subset and was made equivariant to elastic deformations in a self-supervised manner using deformation-aware learning to improve the generalisation performance. The proposed technique was evaluated quantitatively and qualitatively against the test set and achieved a Dice score of 80.44 ± 0.83. Furthermore, the result of the proposed method was compared against a selected manually segmented region (62.07 resultant Dice) and has shown a considerable improvement (18.98%) with deformation-aware learning.

Reorganization of Brain Functional Connectivity Network and Vision Restoration Following Combined tACS-tDCS Treatment After Occipital Stroke.

Objective: Non-invasive brain stimulation (NIBS) is already known to improve visual field functions in patients with optic nerve damage and partially restores the organization of brain functional connectivity networks (FCNs). However, because little is known if NIBS is effective also following brain damage, we now studied the correlation between visual field recovery and FCN reorganization in patients with stroke of the central visual pathway. Method: In a controlled, exploratory trial, 24 patients with hemianopia were randomly assigned to one of three brain stimulation groups: transcranial direct current stimulation (tDCS)/transcranial alternating current stimulation (tACS) (ACDC); sham tDCS/tACS (AC); sham tDCS/sham tACS (Sham), which were compared to age-matched controls (n = 24). Resting-state electroencephalogram (EEG) was collected at baseline, after 10 days stimulation and at 2 months follow-up. EEG recordings were analyzed for FCN measures using graph theory parameters, and FCN small worldness of the network and long pairwise coherence parameter alterations were then correlated with visual field performance. Result: ACDC enhanced alpha-band FCN strength in the superior occipital lobe of the lesioned hemisphere at follow-up. A negative correlation (r = -0.80) was found between the intact visual field size and characteristic path length (CPL) after ACDC with a trend of decreased alpha-band centrality of the intact middle occipital cortex. ACDC also significantly decreased delta band coherence between the lesion and the intact occipital lobe, and coherence was enhanced between occipital and temporal lobe of the intact hemisphere in the low beta band. Responders showed significantly higher strength in the low alpha band at follow-up in the intact lingual and calcarine cortex and in the superior occipital region of the lesioned hemisphere. Conclusion: While ACDC decreases delta band coherence between intact and damaged occipital brain areas indicating inhibition of low-frequency neural oscillations, ACDC increases FCN connectivity between the occipital and temporal lobe in the intact hemisphere. When taken together with the lower global clustering coefficient in responders, these findings suggest that FCN reorganization (here induced by NIBS) is adaptive in stroke. It leads to greater efficiency of neural processing, where the FCN requires fewer connections for visual processing.

Interhemispheric Cortical Network Connectivity Reorganization Predicts Vision Impairment in Stroke.

Stroke is one of the main causes of disability in human beings, and when the occipital lobe is affected, this leads to partial vision loss (homonymous hemianopia). To understand brain mechanisms of vision loss and recovery, graph theory-based brain functional connectivity network (FCN) analysis was recently introduced. However, few brain network studies exist that have studied if the strength of the damaged FCN can predict the extent of functional impairment. We now characterized the brain FCN using deep neural network analysis to describe multiscale brain networks and explore their corresponding physiological patterns. In a group of 24 patients and 24 controls, Bi-directional long short-term memory (Bi-LSTM) was evaluated to reveal the cortical network pattern learning efficiency compared with other traditional algorithms. Bi-LSTM achieved the best balanced-overall accuracy of 73% with sensitivity of 70% and specificity and 75% in the low alpha band. This demonstrates that bi-directional learning can capture the brain network feature representation of both hemispheres. It shows that brain damage leads to reorganized FCN patterns with a greater number of functional connections of intermediate density in the high alpha band. Future studies should explore how this understanding of brain FCN can be used for clinical diagnostics and rehabilitation.

Dual Skip Connections Minimize the False Positive Rate of Lung Nodule Detection in CT images.

Pulmonary cancer is one of the most commonly diagnosed and fatal cancers and is often diagnosed by incidental findings on computed tomography. Automated pulmonary nodule detection is an essential part of computer-aided diagnosis, which is still facing great challenges and difficulties to quickly and accurately locate the exact nodules' positions. This paper proposes a dual skip connection upsampling strategy based on Dual Path network in a U-Net structure generating multiscale feature maps, which aims to minimize the ratio of false positives and maximize the sensitivity for lesion detection of nodules. The results show that our new upsampling strategy improves the performance by having 85.3% sensitivity at 4 FROC per image compared to 84.2% for the regular upsampling strategy or 81.2% for VGG16-based Faster-R-CNN.

Fine-tuning deep learning model parameters for improved super-resolution of dynamic MRI with prior-knowledge.

Dynamic imaging is a beneficial tool for interventions to assess physiological changes. Nonetheless during dynamic MRI, while achieving a high temporal resolution, the spatial resolution is compromised. To overcome this spatio-temporal trade-off, this research presents a super-resolution (SR) MRI reconstruction with prior knowledge based fine-tuning to maximise spatial information while reducing the required scan-time for dynamic MRIs. A U-Net based network with perceptual loss is trained on a benchmark dataset and fine-tuned using one subject-specific static high resolution MRI as prior knowledge to obtain high resolution dynamic images during the inference stage. 3D dynamic data for three subjects were acquired with different parameters to test the generalisation capabilities of the network. The method was tested for different levels of in-plane undersampling for dynamic MRI. The reconstructed dynamic SR results after fine-tuning showed higher similarity with the high resolution ground-truth, while quantitatively achieving statistically significant improvement. The average SSIM of the lowest resolution experimented during this research (6.25% of the k-space) before and after fine-tuning were 0.939 ± 0.008 and 0.957 ± 0.006 respectively. This could theoretically result in an acceleration factor of 16, which can potentially be acquired in less than half a second. The proposed approach shows that the super-resolution MRI reconstruction with prior-information can alleviate the spatio-temporal trade-off in dynamic MRI, even for high acceleration factors.

CHAOS Challenge - combined (CT-MR) healthy abdominal organ segmentation.

Segmentation of abdominal organs has been a comprehensive, yet unresolved, research field for many years. In the last decade, intensive developments in deep learning (DL) introduced new state-of-the-art segmentation systems. Despite outperforming the overall accuracy of existing systems, the effects of DL model properties and parameters on the performance are hard to interpret. This makes comparative analysis a necessary tool towards interpretable studies and systems. Moreover, the performance of DL for emerging learning approaches such as cross-modality and multi-modal semantic segmentation tasks has been rarely discussed. In order to expand the knowledge on these topics, the CHAOS - Combined (CT-MR) Healthy Abdominal Organ Segmentation challenge was organized in conjunction with the IEEE International Symposium on Biomedical Imaging (ISBI), 2019, in Venice, Italy. Abdominal organ segmentation from routine acquisitions plays an important role in several clinical applications, such as pre-surgical planning or morphological and volumetric follow-ups for various diseases. These applications require a certain level of performance on a diverse set of metrics such as maximum symmetric surface distance (MSSD) to determine surgical error-margin or overlap errors for tracking size and shape differences. Previous abdomen related challenges are mainly focused on tumor/lesion detection and/or classification with a single modality. Conversely, CHAOS provides both abdominal CT and MR data from healthy subjects for single and multiple abdominal organ segmentation. Five different but complementary tasks were designed to analyze the capabilities of participating approaches from multiple perspectives. The results were investigated thoroughly, compared with manual annotations and interactive methods. The analysis shows that the performance of DL models for single modality (CT / MR) can show reliable volumetric analysis performance (DICE: 0.98 ± 0.00 / 0.95 ± 0.01), but the best MSSD performance remains limited (21.89 ± 13.94 / 20.85 ± 10.63 mm). The performances of participating models decrease dramatically for cross-modality tasks both for the liver (DICE: 0.88 ± 0.15 MSSD: 36.33 ± 21.97 mm). Despite contrary examples on different applications, multi-tasking DL models designed to segment all organs are observed to perform worse compared to organ-specific ones (performance drop around 5%). Nevertheless, some of the successful models show better performance with their multi-organ versions. We conclude that the exploration of those pros and cons in both single vs multi-organ and cross-modality segmentations is poised to have an impact on further research for developing effective algorithms that would support real-world clinical applications. Finally, having more than 1500 participants and receiving more than 550 submissions, another important contribution of this study is the analysis on shortcomings of challenge organizations such as the effects of multiple submissions and peeking phenomenon.

On the properties of prototype-based fuzzy classifiers.

The use of natural language rules that are able to handle vague and, possibly, even contradicting knowledge in order to model formal dependences is an intriguing idea. Fuzzy IF-THEN rules have been proposed as classification methods that can easily be defined and interpreted by humans or built automatically by learning algorithms. This paper gives an intuitive insight into the properties and the behavior of prototype-based fuzzy classifiers, using formal descriptions and visualization methods. This can help to avoid some common peculiarities and pitfalls in the manual or automated design of fuzzy classifiers.

Visualization and simulation techniques for surgical simulators using actual patient's data.

Because of the increasing complexity of surgical interventions research in surgical simulation became more and more important over the last years. However, the simulation of tissue deformation is still a challenging problem, mainly due to the short response times that are required for real-time interaction. The demands to hard and software are even larger if not only the modeled human anatomy is used but the anatomy of actual patients. This is required if the surgical simulator should be used as training medium for expert surgeons rather than students. In this article, suitable visualization and simulation methods for surgical simulation utilizing actual patient's datasets are described. Therefore, the advantages and disadvantages of direct and indirect volume rendering for the visualization are discussed and a neuro-fuzzy system is described, which can be used for the simulation of interactive tissue deformations. The neuro-fuzzy system makes it possible to define the deformation behavior based on a linguistic description of the tissue characteristics or to learn the dynamics by using measured data of real tissue. Furthermore, a simulator for minimally-invasive neurosurgical interventions is presented that utilizes the described visualization and simulation methods. The structure of the simulator is described in detail and the results of a system evaluation by an experienced neurosurgeon--a quantitative comparison between different methods of virtual endoscopy as well as a comparison between real brain images and virtual endoscopies--are given. The evaluation proved that the simulator provides a higher realism of the visualization and simulation then other currently available simulators.