Correcting electrode modelling errors in EIT on realistic 3D head models

High-resolution conductivity reconstruction by electrical impedance tomography using structure-aware hybrid-fusion learning

BACKGROUND

Electrical impedance tomography (EIT) has gained considerable attention in the medical field for the diagnosis of lung-related diseases, owing to its non-invasive and real-time characteristics. However, due to the ill-posedness and underdetermined nature of the inverse problem in EIT, suboptimal reconstruction performance and reduced robustness against the measurement noise and modeling errors are common issues.

OBJECTIVES

This study aims to mine the deep feature information from measurement voltages, acquired from the EIT sensor, to reconstruct the high-resolution conductivity distribution and enhance the robustness against the measurement noise and modeling errors using the deep learning method.

METHODS

A novel data-driven method named the structure-aware hybrid-fusion learning (SA-HFL) is proposed. SA-HFL is composed of three main components: a segmentation branch, a conductivity reconstruction branch, and a feature fusion module. These branches work in tandem to extract different feature information from the measurement voltage, which is then fused to reconstruct the conductivity distribution. The unique aspect of this network is its ability to utilize different features extracted from various branches to accomplish reconstruction objectives. To supervise the training of the network, we generated regular-shaped and lung-shaped EIT datasets through numerical calculations.

RESULTS

The simulations and three experiments demonstrate that the proposed SA-HFL exhibits superior performance in qualitative and quantitative analyses, compared with five cutting-edge deep learning networks and the optical image-guided group sparsity (IGGS) method. The evaluation metrics, relative error (RE), mean structural similarity index (MSSIM), and peak signal-to-noise ratio (PSNR), are improved by implementing the SA-HFL method. For the regular-shaped dataset, the values are 0.119 (RE), 0.9882 (MSSIM), and 31.03 (PSNR). For the lung-shaped dataset, the values are 0.257 (RE), 0.9151 (MSSIM), and 18.67 (PSNR). Furthermore, the proposed network can be executed with appropriate parameters and efficient floating-point operations per second (FLOPs), concerning network complexity and inference speed.

CONCLUSIONS

The reconstruction results indicate that fusing feature information from different branches enhances the accuracy of conductivity reconstruction in the EIT inverse problem. Moreover, the study shows that fusing different modalities of information to reconstruct the EIT conductivity distribution may be a future development direction.

Applied machine learning for stroke differentiation by electrical impedance tomography with realistic numerical models

Electrical impedance tomography (EIT) may have potential to overcome existing limitations in stroke differentiation, enabling low-cost, rapid, and mobile data collection. Combining bioimpedance measurement technologies such as EIT with machine learning classifiers to support decision-making can avoid commonly faced reconstruction challenges due to the nonlinear and ill-posed nature of EIT imaging. Therefore, in this work, we advance this field through a study integrating realistic head models with clinically relevant test scenarios, and a robust architecture consisting of nested cross-validation and principal component analysis. Specifically, realistic head models are designed which incorporate the highly conductive layers of cerebrospinal fluid in the subarachnoid space and ventricles. In total, 135 unique models are created to represent a large patient population, with normal, haemorrhagic, and ischemic brains. Simulated EIT voltage data generated from these models are used to assess the classification performance of support vector machines. Parameters explored include driving frequency, signal-to-noise ratio, kernel function, and composition of binary classes. Classifier accuracy at 60 dB signal-to-noise ratio, reported as mean and standard deviation, are (79.92% ± 10.82%) for lesion differentiation, (74.78% ± 3.79%) for lesion detection, (77.49% ± 15.90%) for bleed detection, and (60.31% ± 3.98%) for ischemia detection (after ruling out bleed). The results for each method were obtained with statistics from 3 independent runs with 17,280 observations, polynomial kernel functions, and feature reduction of 76% by PCA (from 208 to 50 features). While results of this study show promise for stroke differentiation using EIT data, our findings indicate that the achievable accuracy is highly dependent on the classification scenario and application-specific classifiers may be necessary to achieve acceptable accuracy.

Sensitivity Analysis Highlights the Importance of Accurate Head Models for Electrical Impedance Tomography Monitoring of Intracerebral Hemorrhagic Stroke

OBJECTIVE

Electrical impedance tomography (EIT) has been proposed as a novel tool for diagnosing stroke. However, so far, the clinical feasibility is unresolved. In this study, we aim to investigate the need for accurate head modeling in EIT and how the inhomogeneities of the head contribute to the EIT measurement and affect its feasibility in monitoring the progression of a hemorrhagic stroke.

METHODS

We compared anatomically detailed six- and three-layer finite element models of a human head and computed the resulting scalp electrode potentials and the lead fields of selected electrode configurations. We visualized the resulting EIT measurement sensitivity distributions, computed the scalp electrode potentials, and examined the inverse imaging with selected cases. The effect of accurate tissue geometry and conductivity values on the EIT measurement is assessed with multiple different hemorrhagic perturbation locations and sizes.

RESULTS

Our results show that accurate tissue geometries and conductivity values inside the cranial cavity, especially the highly conductive cerebral spinal fluid, significantly affect EIT measurement sensitivity distribution and measured potentials.

CONCLUSIONS

We can conclude that the three-layer head models commonly used in EIT literature cannot depict the current paths correctly in the head. Thus, our study highlights the need to consider the detailed geometry of the cerebrospinal fluid (CSF) in EIT.

SIGNIFICANCE

The results clearly show that the CSF should be considered in the head EIT calculations.

Self-Abrading Servo Electrode Helmet for Electrical Impedance Tomography

Electrical Impedance Tomography (EIT) is a medical imaging technique which has the potential to reduce time to treatment in acute stroke by rapidly differentiating between ischaemic and haemorrhagic stroke. The potential of these methods has been demonstrated in simulation and phantoms, it has not yet successfully translated to clinical studies, due to high sensitivity to errors in scalp electrode mislocation and poor electrode-skin contact. To overcome these limitations, a novel electrode helmet was designed, bearing 32 independently controlled self-abrading electrodes. The contact impedance was reduced through rotation on an abrasive electrode on the scalp using a combined impedance, rotation and position feedback loop. Potentiometers within each unit measure the electrode tip displacement within 0.1 mm from the rigid helmet body. Characterisation experiments on a large-scale test rig demonstrated that approximately 20 kPa applied pressure and 5 rotations was necessary to achieve the target 5 kΩ contact impedance at 20 Hz. This performance was then replicated in a simplified self-contained unit where spring loaded electrodes are rotated by servo motors. Finally, a 32-channel helmet and controller which sequentially minimised contact impedance and simultaneously located each electrode was built which reduced the electrode application and localisation time to less than five minutes. The results demonstrated the potential of this approach to rapidly apply electrodes in an acute setting, removing a significant barrier for imaging acute stroke with EIT.

Multi-frequency symmetry difference electrical impedance tomography with machine learning for human stroke diagnosis

OBJECTIVE

Multi-frequency symmetry difference electrical impedance tomography (MFSD-EIT) can robustly detect and identify unilateral perturbations in symmetric scenes. Here, an investigation is performed to assess if the algorithm can be successfully applied to identify the aetiology of stroke with the aid of machine learning.

METHODS

Anatomically realistic four-layer finite element method models of the head based on stroke patient images are developed and used to generate EIT data over a 5 Hz-100 Hz frequency range with and without bleed and clot lesions present. Reconstruction generates conductivity maps of each head at each frequency. Application of a quantitative metric assessing changes in symmetry across the sagittal plane of the reconstructed image and over the frequency range allows lesion detection and identification. The algorithm is applied to both simulated and human (n = 34 subjects) data. A classification algorithm is applied to the metric value in order to differentiate between normal, haemorrhage and clot values.

MAIN RESULTS

An average accuracy of 85% is achieved when MFSD-EIT with support vector machines (SVM) classification is used to identify and differentiate bleed from clot in human data, with 77% accuracy when differentiating normal from stroke in human data.

CONCLUSION

Applying a classification algorithm to metrics derived from MFSD-EIT images is a novel and promising technique for detection and identification of perturbations in static scenes.

SIGNIFICANCE

The MFSD-EIT algorithm used with machine learning gives promising results of lesion detection and identification in challenging conditions like stroke. The results imply feasible translation to human patients.

Bi-Frequency Symmetry Difference EIT-Feasibility and Limitations of Application to Stroke Diagnosis

OBJECTIVE

Bi-Frequency Symmetry Difference (BFSD)-EIT can detect, localize and identify unilateral perturbations in symmetric scenes. Here, we test the viability and robustness of BFSD-EIT in stroke diagnosis.

METHODS

A realistic 4-layer Finite Element Method (FEM) head model with and without bleed and clot lesions is developed. Performance is assessed with test parameters including: measurement noise, electrode placement errors, contact impedance errors, deviations in assumed tissue conductivity, deviations in assumed anatomy, and a frequency-dependent background. A final test is performed using ischemic patient data. Results are assessed using images and quantitative metrics.

RESULTS

BFSD-EIT may be feasible for stroke diagnosis if a signal-to-noise ratio (SNR) of ≥60 dB is achievable. Sensitivity to errors in electrode positioning is seen with a tolerance of only ±5 mm, but a tolerance of up to ±30 mm is possible if symmetry is maintained between symmetrically opposite partner electrodes. The technique is robust to errors in contact impedance and assumed tissue conductivity up to at least ±50%. Asymmetric internal anatomy affects performance but may be tolerable for tissues with frequency-dependent conductivity. Errors in assumed external geometry marginally affect performance. A frequency-dependent background does not affect performance with carefully chosen frequency points or use of multiple frequency points across a band. The Global Left-Hand Side (LHS) & Right-Hand Side (RHS) Mean Intensity metric is particularly robust to errors.

CONCLUSION

BFSD-EIT is a promising technique for stroke diagnosis, provided parameters are within the tolerated ranges.

SIGNIFICANCE

BFSD-EIT may prove an important step forward in imaging of static scenes such as stroke.

Brain haemorrhage detection using a SVM classifier with electrical impedance tomography measurement frames

Brain haemorrhages often require urgent treatment with a consequent need for quick and accurate diagnosis. Therefore, in this study, we investigate Support Vector Machine (SVM) classifiers for detecting brain haemorrhages using Electrical Impedance Tomography (EIT) measurement frames. A 2-layer model of the head, along with a series of haemorrhages, is designed as both numerical models and physical phantoms. EIT measurement frames, taken from an electrode array placed on the head surface, are used to train and test linear SVM classifiers. Various scenarios are implemented on both platforms to examine the impact of variables such as noise level, lesion location, lesion size, variation in electrode positioning, and variation in anatomy, on the classifier performance. The classifier performed well in numerical models (sensitivity and specificity of 90%+) with signal-to-noise ratios of 60 dB+, was independent of lesion location, and could detect lesions reliably down to the tested minimum volume of 5 ml. Slight variations in electrode layout did not affect performance. Performance was affected by variations in anatomy however, emphasising the need for large training sets covering different anatomies. The phantom models proved more challenging, with maximal sensitivity and specificity of 75% when used with the linear SVM. Finally, the performance of two more complex classifiers is briefly examined and compared to the linear SVM classifier. These results demonstrate that a radial basis function (RBF) SVM classifier and a neural network classifier can improve detection accuracy. Classifiers applied to EIT measurement frames is a novel approach for lesion detection and may offer an effective diagnostic tool clinically. A challenge is to translate the strong results from numerical models into real world phantoms and ultimately human patients, as well as the selection and development of optimal classifiers for this application.

Reproducible 3D printed head tanks for electrical impedance tomography with realistic shape and conductivity distribution

OBJECTIVE

Electrical impedance tomography (EIT) has many promising applications in brain injury monitoring. To evaluate both instrumentation and reconstruction algorithms, experiments are first performed in head tanks. Existing methods, whilst accurate, produce a discontinuous conductivity, and are often made by hand, making it hard for other researchers to replicate.

APPROACH

We have developed a method for constructing head tanks directly in a 3D printer. Conductivity was controlled through perforations in the skull surface, which allow for saline to pass through. Varying the diameter of the holes allowed for the conductivity to be controlled with 3% error for the target conductivity range. Taking CT and MRI segmentations as a basis, this method was employed to create an adult tank with a continuous conductivity distribution, and a neonatal tank with fontanelles.

MAIN RESULTS

Using 3D scanning a geometric accuracy of 0.21 mm was recorded, equal to that of the precision of the 3D printer used. Differences of 6.1%  ±  6.4% (n  =  11 in 4 tanks) compared to simulations were recorded in c. 800 boundary voltages. This may be attributed to the morphology of the skulls increasing tortuosity effects and hole misalignment. Despite significant differences in errors between three repetitions of the neonatal tank, images of a realistic perturbation could still be reconstructed with different tanks used for the baseline and perturbation datasets.

SIGNIFICANCE

These phantoms can be reproduced by any researcher with access to a 'hobbyist' 3D printer in a matter of days. All design files have been released using an open source license to encourage reproduction and modification.

Correction of electrode modelling errors in multi-frequency EIT imaging

The differentiation of haemorrhagic from ischaemic stroke using electrical impedance tomography (EIT) requires measurements at multiple frequencies, since the general lack of healthy measurements on the same patient excludes time-difference imaging methods. It has previously been shown that the inaccurate modelling of electrodes constitutes one of the largest sources of image artefacts in non-linear multi-frequency EIT applications. To address this issue, we augmented the conductivity Jacobian matrix with a Jacobian matrix with respect to electrode movement. Using this new algorithm, simulated ischaemic and haemorrhagic strokes in a realistic head model were reconstructed for varying degrees of electrode position errors. The simultaneous recovery of conductivity spectra and electrode positions removed most artefacts caused by inaccurately modelled electrodes. Reconstructions were stable for electrode position errors of up to 1.5 mm standard deviation along both surface dimensions. We conclude that this method can be used for electrode model correction in multi-frequency EIT.