Home » Posts tagged 'Signal Processing'
Tag Archives: Signal Processing
Su-081-Spatial global component filter vs short channel regression for removing non-cortical component of fNIRS
Non-cortical components can be prominent in fNIRS recordings. These components can be related to blood pressure or respiratory activity and should be removed to measure reliable neural activity or functional connectivity. Here we compare two methods for removing non-cortical components, spatial global component filtering (GCR) and short channel regression (SCR). SCR requires short channel signal acquisition with high signal-to-noise ratio (SNR) and only SNR/(1+SNR) of non-cortical component can be removed. GCR requires large channel coverage and will reduce activity that has large spatial extent. Both methods should be used in appropriate manner. Continue reading
Fr-109-Individualized head models from registered optodes or photogrammetry for improved DOT image reconstruction
We propose a data-driven algorithm to approximate individual head anatomies in case of unavailable MRI/CT scans to improve image reconstruction accuracy over widely used standard head models such as Colin27. Based on a low-dimensional representation of a large head model database we derive individual head parameters solely from additional knowledge of the subject’s scalp as gained from e.g. photogrammetry scans or exact optode positions. Preliminary photon simulation results show significantly less deviation from the subject’s ground truth head model compared to Colin27 and raises hopes for improved source reconstructions with little effort in many future fNIRS studies. Continue reading
Sa-018-A Framework for Synthetic fNIRS Data Generation
Synthetic data generation is a necessary mechanism for validation efforts requiring a compromise between the simplicity of the forward modelling and the required modelling constraints, and a large number of ad-hoc options are available to whoever needs to generate this data. We exploit the bilinear model for flexible fNIRS data emulation. Our framework provides a range of noise functions that integrate both fully synthetic and semisynthetic physiological and instrumental noise through a coherent interface. This framework facilitates synthetic data generation with implications for algorithm validation and enhancing the quality of pipelines. Continue reading
Sa-032-An infant within-subject multimodal NIRS-EEG classifier
In this work we present a within-subject multivariate pattern analysis using concurrent fNIRS and EEG measures to discriminate patterns elicited by different auditory conditions in newborns. Continue reading
Sa-071-An augmented reality guided optode positioning system on a head-mounted display device
The optode positioning system using augmented reality (AR) was implemented in a head-mounted display (HMD) device. The system provides an AR representation of computer graphics (CGs) such as a brain structure and positional indicators just above the brain regions of interest (ROIs) in virtual space, superimposed on the subject's head in real space. The operator wearing the HMD device can attach the optodes to the subject using both hands freely while checking the location of the ROIs with the CGs. In the preliminary experiment with the head mannequin, it took approximately 3 minutes from system launch to completion of the optode attachment. Continue reading
Sa-084-Accounting for head size distributions in cross-sectional and longitudinal pediatric fNIRS studies
fNIRS is sensitive to the structure and thickness of extra-cerebral layers— particularly, the sinus cavities and the layer of cerebral spinal fluid (CSF). This uncertainty creates the potential for systematic bias in analysis of fNIRS measurements in longitudinal development studies, because the sensitivity of fNIRS measurements decays exponentially with the depth of brain activity from the surface sensors. The study characterizes differences in head anatomy and their effect on expected fNIRS sensitivity using existing structural MRI data from pediatric populations, then develops and compares analytical methods to account for this variability in fNIRS analysis focusing on cross-sectional and longitudinal studies. Continue reading
Sa-082-Effects of Matrix Conditioning strategies on Multifrequency High-density Diffuse Optical Tomography
Kindly input your 100-word abstract below (please remove the current text beforehand). This abstract must also be presented in the accompanying 1-page summary PDF attachment. We strongly advise reviewing all submission requirements and guidelines at https://fnirs2024.fnirs.org/cfp/ before submitting your abstract. For your convenience, the template for the 1-page summary can be downloaded from http://fnirs2024.fnirs.org/wp-content/uploads/2024/01/fNIRS2024-abstract-template_v001.docx. Including the names of all contributing authors in the format for abstract submission is crucial. Please note that no changes to authorship will be accepted after the submission has been made. If you have any questions or experience any issues, please contact Augusto Bonilauri. Continue reading
Sa-098-Artefact detection and removal using ICA-ERBM in fNIRS
We introduce a blind source separation technique for the detection and removal of artefacts in fNIRS data. Synthetic HRFs are added to optical density timeseries recorded from three subjects with simultaneous accelerometer and gyroscope recordings. ICA-ERBM is used to identify and remove artefacts correlated with motion. We compare the recovery of the synthetic HRFs after motion correction across methods and observe strong performance from the ICA-ERBM algorithm indicating, that with further hyperparamter tuning, it has potential to effectively clean fNIRS signals. Continue reading
Su-012-Comparison of Functional Intraoperative Optical and Pre-operative Magnetic Resonance Imaging in Resting-state and Task-based Procedures
Functional Magnetic Resonance Imaging (fMRI) and Functional Optical Imaging (fRGB) are method of identifying functional brain region, preoperatively and intraoperatively, respectively. Each method contains two procedural modes: task-based and resting-state. Unlike task-based imaging, resting-state imaging does not require patient intervention which may facilitate functional identification during neurosurgery. Thus, we evaluated the capability of identifying brain function using intraoperative resting-state fRGB by comparing the functional areas to those identified using fMRI. For nine patients with gliomas, the fMRI data was projected onto the fRGB optical space, and we compared the activated regions using DICE and overlap coefficients. Continue reading
Su-008-Immediate Measurement of Placental Depth During Time Domain NIRS using Deep Learning
The reliability of NIRS devices for placental oxygenation monitoring can be affected by anatomical variation. Although the depth of the placenta can be obtained using a secondary modality like ultrasound to warn that a large depth may undermine the result, this is often not available. Here, we show that the raw time point spread function from a TD-NIRS system can be used to predict the placental depth through a deep learning model. As a relatively small amount of raw training data was needed, the approach could be extended to other analysis applications. Continue reading