Comparison of source localization techniques in diffuse optical tomography for fNIRS application using a realistic head model
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Functional near-infrared spectroscopy (fNIRS) is a non-invasive imaging technique that elicits growing interest for research and clinical applications. In the last decade, efforts have been made to develop a mathematical framework in order to image the effective sources of hemoglobin variations in brain tissues. Different approaches can be used to impose additional information or constraints when reconstructing the cerebral images of an ill-posed problem. The goal of this study is to compare the performance and limitations of several source localization techniques in the context of fNIRS tomography using individual anatomical magnetic resonance imaging (MRI) to model light propagation. The forward problem is solved using a Monte Carlo simulation of light propagation in the tissues. The inverse problem has been linearized using the Rytov approximation. Then, Tikhonov regularization applied to least squares, truncated singular value decomposition, back-projection, L1-norm regularization, minimum norm estimates, low resolution electromagnetic tomography and Bayesian model averaging techniques are compared using a receiver operating characteristic analysis, blurring and localization error measures. Using realistic simulations (n = 450) and data acquired from a human participant, this study depicts how these source localization techniques behave in a human head fNIRS tomography. When compared to other methods, Bayesian model averaging is proposed as a promising method in DOT and shows great potential to improve specificity, accuracy, as well as to reduce blurring and localization error even in presence of noise and deep sources. Classical reconstruction methods, such as regularized least squares, offer better sensitivity but higher blurring; while more novel L1-based method provides sparse solutions with small blurring and high specificity but lower sensitivity. The application of these methods is also demonstrated experimentally using visual fNIRS experiment with adult participant.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it