Deep Pinsker and James-Stein Neural Networks for Decoding Motor Intentions From Limited Data.

Non-parametric regression has been shown to be useful in extracting relevant features from Local Field Potential (LFP) signals for decoding motor intentions. Yet, in many instances, brain-computer interfaces (BCIs) rely on simple classification methods, circumventing deep neural networks (DNNs) due to limited training data. This paper leverages the robustness of several important results in non-parametric regression to harness the potentials of deep learning in limited data setups. We consider a solution that combines Pinsker's theorem as well as its adaptively optimal counterpart due to James-Stein for feature extraction from LFPs, followed by a DNN for classifying motor intentions. We apply our approach to the problem of decoding eye movement intentions from LFPs collected in macaque cortex while the animals perform memory-guided visual saccades to one of eight target locations. The results demonstrate that a DNN classifier trained over the Pinsker features outperforms the benchmark method based on linear discriminant analysis (LDA) trained over the same features.

Cross-subject decoding of eye movement goals from local field potentials.

OBJECTIVE: We consider the cross-subject decoding problem from local field potential (LFP) signals, where training data collected from the prefrontal cortex (PFC) of a source subject is used to decode intended motor actions in a destination subject. APPROACH: We propose a novel supervised transfer learning technique, referred to as data centering, which is used to adapt the feature space of the source to the feature space of the destination. The key ingredients of data centering are the transfer functions used to model the deterministic component of the relationship between the source and destination feature spaces. We propose an efficient data-driven estimation approach for linear transfer functions that uses the first and second order moments of the class-conditional distributions. MAIN RESULTS: We apply our data centering technique with linear transfer functions for cross-subject decoding of eye movement intentions in an experiment where two macaque monkeys perform memory-guided visual saccades to one of eight target locations. The results show peak cross-subject decoding performance of [Formula: see text], which marks a substantial improvement over random choice decoder. In addition to this, data centering also outperforms standard sampling-based methods in setups with imbalanced training data. SIGNIFICANCE: The analyses presented herein demonstrate that the proposed data centering is a viable novel technique for reliable LFP-based cross-subject brain-computer interfacing and neural prostheses.

Sparse Representation-Based Denoising for High-Resolution Brain Activation and Functional Connectivity Modeling: A Task fMRI Study.

In the field of neuroimaging and cognitive neuroscience, functional Magnetic Resonance Imaging (fMRI) has been widely used to study the functional localization and connectivity of the brain. However, the inherently low signal-to-noise ratio (SNR) of the fMRI signals greatly limits the accuracy and resolution of current studies. In addressing this fundamental challenge in fMRI analytics, in this work we develop and implement a denoising method for task fMRI (tfMRI) data in order to delineate the high-resolution spatial pattern of the brain activation and functional connectivity via dictionary learning and sparse coding (DLSC). In addition to the traditional unsupervised dictionary learning model which has shown success in image denoising, we further utilize the prior knowledge of task paradigm to learn a dictionary consisting of both data-driven and model-driven terms for a more stable sparse representation of the data. The proposed method is applied to preprocess the motor tfMRI dataset from Human Connectome Project (HCP) for the purpose of brain activation detection and functional connectivity estimation. Comparison between the results from original and denoised fMRI data shows that the disruptive brain activation and functional connectivity patterns can be recovered, and the prominence of such patterns is improved through denoising. The proposed method is then compared with the temporal non-local means (tNLM)-based denoising method and shows consistently superior performance in various experimental settings. The promising results show that the proposed DLSC-based fMRI denoising method can effectively reduce the noise level of the fMRI signals and increase the interpretability of the inferred results, therefore constituting a crucial part of the preprocessing pipeline and provide the foundation for further high-resolution functional analysis.

Minimax-optimal decoding of movement goals from local field potentials using complex spectral features.

OBJECTIVE: We consider the problem of predicting eye movement goals from local field potentials (LFP) recorded through a multielectrode array in the macaque prefrontal cortex. The monkey is tasked with performing memory-guided saccades to one of eight targets during which LFP activity is recorded and used to train a decoder. APPROACH: Previous reports have mainly relied on the spectral amplitude of the LFPs as decoding feature, while neglecting the phase without proper theoretical justification. This paper formulates the problem of decoding eye movement intentions in a statistically optimal framework and uses Gaussian sequence modeling and Pinsker's theorem to generate minimax-optimal estimates of the LFP signals which are used as decoding features. The approach is shown to act as a low-pass filter and each LFP in the feature space is represented via its complex Fourier coefficients after appropriate shrinking such that higher frequency components are attenuated; this way, the phase information inherently present in the LFP signal is naturally embedded into the feature space. MAIN RESULTS: We show that the proposed complex spectrum-based decoder achieves prediction accuracy of up to [Formula: see text] at superficial cortical depths near the surface of the prefrontal cortex; this marks a significant performance improvement over conventional power spectrum-based decoders. SIGNIFICANCE: The presented analyses showcase the promising potential of low-pass filtered LFP signals for highly reliable neural decoding of intended motor actions.

Complementary lattice arrays for coded aperture imaging.

In this work, we propose the concept of complementary lattice arrays in order to enable a broader range of designs for coded aperture imaging systems. We provide a general framework and methods that generate richer and more flexible designs compared to the existing techniques. Besides this, we review and interpret the state-of-the-art uniformly redundant array designs, broaden the related concepts, and propose new design methods.

In vivo and in situ measurement and modelling of intra-body effective complex permittivity.

Radio frequency tracking of medical micro-robots in minimally invasive medicine is usually investigated upon the assumption that the human body is a homogeneous propagation medium. In this Letter, the authors conducted various trial programs to measure and model the effective complex permittivity ε in terms of refraction ε', absorption ε″ and their variations in gastrointestinal (GI) tract organs (i.e. oesophagus, stomach, small intestine and large intestine) and the porcine abdominal wall under in vivo and in situ conditions. They further investigated the effects of irregular and unsynchronised contractions and simulated peristaltic movements of the GI tract organs inside the abdominal cavity and in the presence of the abdominal wall on the measurements and variations of ε' and ε''. They advanced the previous models of effective complex permittivity of a multilayer inhomogeneous medium, by estimating an analytical model that accounts for reflections between the layers and calculates the attenuation that the wave encounters as it traverses the GI tract and the abdominal wall. They observed that deviation from the specified nominal layer thicknesses due to non-geometric boundaries of GI tract morphometric variables has an impact on the performance of the authors' model. Therefore, they derived statistical-based models for ε' and ε'' using their experimental measurements.

Sparse Signal Recovery from a Mixture of Linear and Magnitude-Only Measurements.

We consider the problem of exact sparse signal recovery from a combination of linear and magnitude-only (phaseless) measurements. A k-sparse signal x ∈ ℂ n is measured as r = Bx and y = |Cx|, where B ∈ ℂ m1×n and C ∈ ℂ m2×n are measurement matrices and | · | is the element-wise absolute value. We show that if max(2m1, 1) + m2 ≥ 4k - 1, then a set of generic measurements are sufficient to recover every k-sparse x exactly, establishing the trade-off between the number of linear and magnitude-only measurements.

An augmented Lagrangian based compressed sensing reconstruction for non-Cartesian magnetic resonance imaging without gridding and regridding at every iteration.

BACKGROUND: Non-Cartesian trajectories are used in a variety of fast imaging applications, due to the incoherent image domain artifacts they create when undersampled. While the gridding technique is commonly utilized for reconstruction, the incoherent artifacts may be further removed using compressed sensing (CS). CS reconstruction is typically done using conjugate-gradient (CG) type algorithms, which require gridding and regridding to be performed at every iteration. This leads to a large computational overhead that hinders its applicability. METHODS: We sought to develop an alternative method for CS reconstruction that only requires two gridding and one regridding operation in total, irrespective of the number of iterations. This proposed technique is evaluated on phantom images and whole-heart coronary MRI acquired using 3D radial trajectories, and compared to conventional CS reconstruction using CG algorithms in terms of quantitative vessel sharpness, vessel length, computation time, and convergence rate. RESULTS: Both CS reconstructions result in similar vessel length (P = 0.30) and vessel sharpness (P = 0.62). The per-iteration complexity of the proposed technique is approximately 3-fold lower than the conventional CS reconstruction (17.55 vs. 52.48 seconds in C++). Furthermore, for in-vivo datasets, the convergence rate of the proposed technique is faster (60±13 vs. 455±320 iterations) leading to a ∼23-fold reduction in reconstruction time. CONCLUSIONS: The proposed reconstruction provides images of similar quality to the conventional CS technique in terms of removing artifacts, but at a much lower computational complexity.

Compressed sensing reconstruction for undersampled breath-hold radial cine imaging with auxiliary free-breathing data.

PURPOSE: To improve compressed sensing (CS) reconstruction of accelerated breath-hold (BH) radial cine magnetic resonance imaging (MRI) by exploiting auxiliary data acquired between different BHs. MATERIALS AND METHODS: Cardiac function is usually assessed using segmented cine acquisitions over multiple BHs to cover the entire left ventricle (LV). Subjects are given a resting period between adjacent BHs, when conventionally no data are acquired and subjects rest in the scanner. In this study the resting periods between BHs were used to acquire additional free-breathing (FB) data, which are subsequently used to generate a sparsity constraint for each cardiac phase. Images reconstructed using the proposed sparsity constraint were compared with conventional CS using a composite image generated by averaging different cardiac phases. The efficacy of the proposed reconstruction was compared using indices of LV function and blood-myocardium sharpness. RESULTS: The proposed method provided accurate LV ejection fraction measurements for 33% and 20% sampled datasets compared with fully sampled reference images, and showed 14% and 11% higher blood-myocardium border sharpness scores compared to the conventional CS. CONCLUSION: The FB data acquired during resting periods can be efficiently used to improve the image quality of the undersampled BH data without increasing the total scan time.

Bayesian-based localization of wireless capsule endoscope using received signal strength.

In wireless body area sensor networking (WBASN) applications such as gastrointestinal (GI) tract monitoring using wireless video capsule endoscopy (WCE), the performance of out-of-body wireless link propagating through different body media (i.e. blood, fat, muscle and bone) is still under investigation. Most of the localization algorithms are vulnerable to the variations of path-loss coefficient resulting in unreliable location estimation. In this paper, we propose a novel robust probabilistic Bayesian-based approach using received-signal-strength (RSS) measurements that accounts for Rayleigh fading, variable path-loss exponent and uncertainty in location information received from the neighboring nodes and anchors. The results of this study showed that the localization root mean square error of our Bayesian-based method was 1.6 mm which was very close to the optimum Cramer-Rao lower bound (CRLB) and significantly smaller than that of other existing localization approaches (i.e. classical MDS (64.2mm), dwMDS (32.2mm), MLE (36.3mm) and POCS (2.3mm)).

Accelerated aortic flow assessment with compressed sensing with and without use of the sparsity of the complex difference image.

Phase contrast (PC) cardiac MR is widely used for the clinical assessment of blood flow in cardiovascular disease. One of the challenges of PC cardiac MR is the long scan time which limits both spatial and temporal resolution. Compressed sensing reconstruction with accelerated PC acquisitions is a promising technique to increase the scan efficiency. In this study, we sought to use the sparsity of the complex difference of the two flow-encoded images as an additional constraint term to improve the compressed sensing reconstruction of the corresponding accelerated PC data acquisition. Using retrospectively under-sampled data, the proposed reconstruction technique was optimized and validated in vivo on 15 healthy subjects. Then, prospectively under-sampled data was acquired on 11 healthy subjects and reconstructed with the proposed technique. The results show that there is good agreement between the cardiac output measurements from the fully sampled data and the proposed compressed sensing reconstruction method using complex difference sparsity up to acceleration rate 5. In conclusion, we have developed and evaluated an improved reconstruction technique for accelerated PC cardiac MR that uses the sparsity of the complex difference of the two flow-encoded images.

Compressed sensing reconstruction for whole-heart imaging with 3D radial trajectories: a graphics processing unit implementation.

A disadvantage of three-dimensional (3D) isotropic acquisition in whole-heart coronary MRI is the prolonged data acquisition time. Isotropic 3D radial trajectories allow undersampling of k-space data in all three spatial dimensions, enabling accelerated acquisition of the volumetric data. Compressed sensing (CS) reconstruction can provide further acceleration in the acquisition by removing the incoherent artifacts due to undersampling and improving the image quality. However, the heavy computational overhead of the CS reconstruction has been a limiting factor for its application. In this article, a parallelized implementation of an iterative CS reconstruction method for 3D radial acquisitions using a commercial graphics processing unit is presented. The execution time of the graphics processing unit-implemented CS reconstruction was compared with that of the C++ implementation, and the efficacy of the undersampled 3D radial acquisition with CS reconstruction was investigated in both phantom and whole-heart coronary data sets. Subsequently, the efficacy of CS in suppressing streaking artifacts in 3D whole-heart coronary MRI with 3D radial imaging and its convergence properties were studied. The CS reconstruction provides improved image quality (in terms of vessel sharpness and suppression of noise-like artifacts) compared with the conventional 3D gridding algorithm, and the graphics processing unit implementation greatly reduces the execution time of CS reconstruction yielding 34-54 times speed-up compared with C++ implementation.

DiBa: a data-driven Bayesian algorithm for sleep spindle detection.

Although the spontaneous brain rhythms of sleep have commanded much recent interest, their detection and analysis remains suboptimal. In this paper, we develop a data-driven Bayesian algorithm for sleep spindle detection on the electroencephalography (EEG). The algorithm exploits the Karhunen-Loève transform and Bayesian hypothesis testing to produce the instantaneous probability of a spindle's presence with maximal resolution. In addition to possessing flexibility, transparency, and scalability, this algorithm could perform at levels superior to standard methods for EEG event detection.

Compressed-sensing motion compensation (CosMo): a joint prospective-retrospective respiratory navigator for coronary MRI.

Prospective right hemidiaphragm navigator (NAV) is commonly used in free-breathing coronary MRI. The NAV results in an increase in acquisition time to allow for resampling of the motion-corrupted k-space data. In this study, we are presenting a joint prospective-retrospective NAV motion compensation algorithm called compressed-sensing motion compensation (CosMo). The inner k-space region is acquired using a prospective NAV; for the outer k-space, a NAV is only used to reject the motion-corrupted data without reacquiring them. Subsequently, those unfilled k-space lines are retrospectively estimated using compressed sensing reconstruction. We imaged right coronary artery in nine healthy adult subjects. An undersampling probability map and sidelobe-to-peak ratio were calculated to study the pattern of undersampling, generated by NAV. Right coronary artery images were then retrospectively reconstructed using compressed-sensing motion compensation for gating windows between 3 and 10 mm and compared with the ones fully acquired within the gating windows. Qualitative imaging score and quantitative vessel sharpness were calculated for each reconstruction. The probability map and sidelobe-to-peak ratio show that the NAV generates a random undersampling k-space pattern. There were no statistically significant differences between the vessel sharpness and subjective score of the two reconstructions. Compressed-sensing motion compensation could be an alternative motion compensation technique for free-breathing coronary MRI that can be used to reduce scan time.

Compressed sensing with wavelet domain dependencies for coronary MRI: a retrospective study.

Coronary magnetic resonance imaging (MRI) is a noninvasive imaging modality for diagnosis of coronary artery disease. One of the limitations of coronary MRI is its long acquisition time due to the need of imaging with high spatial resolution and constraints on respiratory and cardiac motions. Compressed sensing (CS) has been recently utilized to accelerate image acquisition in MRI. In this paper, we develop an improved CS reconstruction method, Bayesian least squares-Gaussian scale mixture (BLS-GSM), that uses dependencies of wavelet domain coefficients to reduce the observed blurring and reconstruction artifacts in coronary MRI using traditional l(1) regularization. Images of left and right coronary MRI was acquired in 7 healthy subjects with fully-sampled k-space data. The data was retrospectively undersampled using acceleration rates of 2, 4, 6, and 8 and reconstructed using l(1) thresholding, l(1) minimization and BLS-GSM thresholding. Reconstructed right and left coronary images were compared with fully-sampled reconstructions in vessel sharpness and subjective image quality (1-4 for poor-excellent). Mean square error (MSE) was also calculated for each reconstruction. There were no significant differences between the fully sampled image score versus rate 2, 4, or 6 for BLS-GSM for both right and left coronaries (=N.S.). However, for l(1) thresholding significant differences were observed for rates higher than 2 and 4 for right and left coronaries respectively. l(1) minimization also yields images with lower scores compared to the reference for rates higher than 4 for both coronaries. These results were consistent with the quantitative vessel sharpness readings. BLS-GSM allows acceleration of coronary MRI with acceleration rates beyond what can be achieved with l(1) regularization.

Accelerated noncontrast-enhanced pulmonary vein MRA with distributed compressed sensing.

PURPOSE: To investigate the efficacy of distributed compressed sensing (CS) to accelerate free-breathing, electrocardiogram (ECG)-triggered noncontrast pulmonary vein (PV) magnetic resonance angiography (MRA). MATERIALS AND METHODS: Fully sampled ECG-triggered noncontrast PV MRA, using a spatially selective slab inversion preparation sequence, was acquired on seven healthy adult subjects (27 ± 17 years, range: 19-65 years, 4 women). The k-space data were retrospectively randomly undersampled by factors of 2, 4, 6, 8, and 10 and then reconstructed using distributed CS and coil-by-coil CS methods. The reconstructed images were evaluated by two blinded readers in consensus for assessment of major PV branches as well as the presence of artifacts in left atrium (LA) and elsewhere. Diameters of right inferior and right superior PV branches were measured. Additionally, mean square errors (MSE) of the reconstructions were calculated. RESULTS: Both CS methods resulted in image quality scores similar to the fully sampled reference images at undersampling factors up to 6-fold for distributed CS and 4-fold for coil-by-coil CS reconstructions. There was no difference in the presence of artifacts in LA and freedom from important artifacts elsewhere between the two techniques up to undersampling factors of 10 compared to the fully sampled reconstruction. For the PV diameters, no systematic variation between the reference and the reconstructions were observed for either technique. There were no significant differences in MSE between the two methods when compared at a given rate, but the difference was significant when compared across all rates. CONCLUSION: The sparsity of noncontrast PV MRA and the joint sparsity of different coil images allow imaging at high undersampling factors (up to 6-fold) when distributed CS is used.

Low-dimensional-structure self-learning and thresholding: regularization beyond compressed sensing for MRI reconstruction.

An improved image reconstruction method from undersampled k-space data, low-dimensional-structure self-learning and thresholding (LOST), which utilizes the structure from the underlying image is presented. A low-resolution image from the fully sampled k-space center is reconstructed to learn image patches of similar anatomical characteristics. These patches are arranged into "similarity clusters," which are subsequently processed for dealiasing and artifact removal, using underlying low-dimensional properties. The efficacy of the proposed method in scan time reduction was assessed in a pilot coronary MRI study. Initially, in a retrospective study on 10 healthy adult subjects, we evaluated retrospective undersampling and reconstruction using LOST, wavelet-based l(1)-norm minimization, and total variation compressed sensing. Quantitative measures of vessel sharpness and mean square error, and qualitative image scores were used to compare reconstruction for rates of 2, 3, and 4. Subsequently, in a prospective study, coronary MRI data were acquired using these rates, and LOST-reconstructed images were compared with an accelerated data acquisition using uniform undersampling and sensitivity encoding reconstruction. Subjective image quality and sharpness data indicate that LOST outperforms the alternative techniques for all rates. The prospective LOST yields images with superior quality compared with sensitivity encoding or l(1)-minimization compressed sensing. The proposed LOST technique greatly improves image reconstruction for accelerated coronary MRI acquisitions.