Pith. sign in

REVIEW 78 references

Bayesian Model Parameter Learning in Linear Inverse Problems: Application in EEG Focal Source Imaging

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2501.13109 v2 pith:D4E4IDFJ submitted 2025-01-07 eess.SP stat.ML

Alexandra Koulouri , Ville Rimpilainen This is my paper · ORCID
classification eess.SPstat.ML
keywords modelerrorparametersignallearningunknownapproachapproximation
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Inverse problems can be described as limited-data problems in which the signal of interest cannot be observed directly. A physics-based forward model that relates the signal with the observations is typically needed. Unfortunately, unknown model parameters and imperfect forward models can undermine the signal recovery. Even though supervised machine learning offers promising avenues to improve the robustness of the solutions, we have to rely on model-based learning when there is no access to ground truth for the training. Here, we studied a linear inverse problem that included an unknown non-linear model parameter and utilized a Bayesian model-based learning approach that allowed signal recovery and subsequently estimation of the model parameter. This approach, called Bayesian Approximation Error approach, employed a simplified model of the physics of the problem augmented with an approximation error term that compensated for the simplification. An error subspace was spanned with the help of the eigenvectors of the approximation error covariance matrix which allowed, alongside the primary signal, simultaneous estimation of the induced error. The estimated error and signal were then used to determine the unknown model parameter. For the model parameter estimation, we tested different approaches: a conditional Gaussian regression, an iterative (model-based) optimization, and a Gaussian process that was modeled with the help of physics-informed learning. In addition, alternating optimization was used as a reference method. As an example application, we focused on the problem of reconstructing brain activity from EEG recordings under the condition that the electrical conductivity of the patient's skull was unknown in the model. Our results demonstrated clear improvements in EEG source localization accuracy and provided feasible estimates for the unknown model parameter, skull conductivity.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

78 extracted references · 61 canonical work pages

  1. [1]

    Kirsch A 1996 An introduction to the mathematical theory of inverse problems (NY, USA: Springer- Verlag New York.) ISBN 0-387-94530-X

    1996

  2. [2]

    Tikhonov A 1943 Doklady Akademii nauk SSSR 39 195–198

    1943

  3. [3]

    Plato R and Vainikko G 1990 Numerische Mathematik 57(1) 63–79

    1990

  4. [4]

    Engl H W, Hanke M and Neubauer A 1996 Regularization of Inverse Problems (Kluwer Academic Publishers)

    1996

  5. [5]

    Hansen P C 1998 Rank-deficient and Discrete Ill-posed Problems: Numerical Aspects of Linear Inversion (Philadelphia, PA, USA: Society for Industrial and Applied Mathematics) ISBN 0- 89871-403-6

    1998

  6. [6]

    Benning M and Burger M 2018 Acta Numerica 27 1–111 ISSN 1474-0508

    2018

  7. [7]

    Afkham B M, Chung J and Chung M 2021 Inverse Problems 37 105017 ISSN 1361-6420

    2021

  8. [8]

    Mattsson P, Zachariah D and Stoica P 2023 IEEE Transactions on Signal Processing 71 1175–1183 ISSN 1941-0476

    2023

Show all 78 references
  1. [9]

    Burger M and Kabri S 2013 arXiv (Preprint 2312.09845)

    2013 arXiv

  2. [10]

    Campisi P and Egiazarian K (eds) 2007 Blind Image Deconvolution: Theory and Applications (Boca Raton, FL: CRC Press) ISBN 9780849390725

    2007

  3. [11]

    Ni Y and Strohmer T 2024 Auto-calibration and biconvex compressive sensing with applications to parallel mri

    2024

  4. [12]

    Ahmed A, Recht B and Romberg J 2014 IEEE Transactions on Information Theory 60 1711–1732 ISSN 1557-9654

    2014

  5. [13]

    Ling S and Strohmer T 2015 Inverse Problems 31 115002 ISSN 1361-6420

    2015

  6. [14]

    Bolte J, Combettes P L and Pesquet J C 2010 Alternating proximal algorithm for blind image recovery 2010 IEEE International Conference on Image Processing (IEEE)

    2010

  7. [15]

    Li X, Ling S, Strohmer T and Wei K 2019 Applied and Computational Harmonic Analysis 47 893–934 ISSN 1063-5203

    2019

  8. [16]

    Adler J and ¨Oktem O 2017 Inverse Problems 33 124007 ISSN 1361-6420

    2017

  9. [17]

    Adler J and Oktem O 2018 IEEE Transactions on Medical Imaging 37 1322–1332 ISSN 1558-254X

    2018

  10. [18]

    Jin K H, McCann M T, Froustey E and Unser M 2017 IEEE Transactions on Image Processing 26 4509–4522 ISSN 1941-0042

    2017

  11. [19]

    Arridge S, Maass P, ¨Oktem O and Sch¨ onlieb C B 2019Acta Numerica 28 1–174 ISSN 1474-0508

  12. [20]

    Lunz S, Hauptmann A, Tarvainen T, Sch¨ onlieb C B and Arridge S 2021SIAM Journal on Imaging Sciences 14 92–127 ISSN 1936-4954

    1936

  13. [21]

    Arridge S, Hauptmann A and Korolev Y 2023 Inverse problems with learned forward operators

    2023

  14. [22]

    LeCun Y, Bengio Y and Hinton G 2015 Nature 521 436–444 ISSN 1476-4687

    2015

  15. [23]

    Murphy K P 2022 Probabilistic Machine Learning (MIT Press Ltd) ISBN 0262046822 URL https: //www.ebook.de/de/product/41791478/kevin p murphy probabilistic machine learning.html

    2022

  16. [24]

    Kaipio J P and Somersalo E 2004 Statistical and Computational Inverse Problems Applied Mathematical Series (Springer)

    2004

  17. [25]

    Kaipio J and Somersalo E 2007 J. Comput. Appl. Math. 198 493–504

    2007

  18. [26]

    Vanrumste B, Hoey G V, de Walle R V, D’Have M, Lemahieu I and Boon P 2000 Med. Biol. Eng. Comput. 38 528–534

    2000

  19. [27]

    Brain Mapp

    Dannhauer M, Lanfer B, Wolters C and Kn¨ osche T 2011 Hum. Brain Mapp. 32 1383–1399

    2011

  20. [28]

    Lew S, Sliva D D, Choe M, Grant P E, Okada Y, Wolters C H and H¨ am¨ al¨ ainen M S 2013 NeuroImage 76 282–293

    2013

  21. [29]

    Ollikainen J O, Vauhkonen M, Karjalainen P A and Kaipio J P 1999 Med. Eng. Phys. 21(3) 143–154

    1999

  22. [30]

    27(1) 95–111

    Montes-Restrepo V, van Mierlo P, Strobbe G, Staelens S, Vanderberghe S and Hallez H 2014 Brain Topogr. 27(1) 95–111

    2014

  23. [31]

    Rimpilainen V, Koulouri A, Lucka F, Kaipio J P and Wolters C H 2019 NeuroImage 188 252–260 35

    2019

  24. [32]

    Vorwerk J, Aydin U, Wolters C H and Butson C R 2019 Frontiers in Neuroscience 13 ISSN 1662-453X

    2019

  25. [33]

    Lew S, Wolters C H, Anwander A, Makeig S and MacLeod R S 2009 Human Brain Mapping 30 2862–2878

    2009

  26. [34]

    Papageorgakis C 2017 Patient specific conductivity models: characterization of the skull bones. Ph.D. thesis

    2017

  27. [36]

    Huang M X, Song T, Jr D J H, Podgorny I, Jousmaki V, Cui L, Harrington D L, Dale A M, Lee R R, Elman J and Halgren E 2007 NeuroImage 37 731–748

    2007

  28. [37]

    Aydin U, Vorwerk J, K¨ upper P, Heers M, Kugel H, Galka A, Hamid L, Wellmer J, Kellinghaus C, Rampp S and Wolters C H 2014 PLoS ONE 9(3) e93154

    2014

  29. [38]

    Antonakakis M, Schrader S, Aydin U, Khan A, Gross J, Zervakis M, Rampp S and Wolters C H 2020 NeuroImage 223 117353 ISSN 1053-8119

    2020

  30. [39]

    Schmidt C, Wagner S, Burger M, Rienen U v and Wolters C H 2015 Journal of Neural Engineering 12 046028 ISSN 1741-2552

    2015

  31. [40]

    Saturnino G B, Thielscher A, Madsen K H, Kn¨ osche T R and Weise K 2019 NeuroIm- age 188 821–834 ISSN 1053-8119 URL https://www.sciencedirect.com/science/article/pii/ S1053811918322031

    2019

  32. [41]

    McCann H and Beltrachini L 2021 Biomedical Physics & Engineering Express 7 045018 URL https://dx.doi.org/10.1088/2057-1976/ac0547

    2021 doi

  33. [42]

    Kaipio J and Kolehmainen V 2013 Approximate marginalization over modeling errors and uncertainties in inverse problems Bayesian Theory and Applications ed Damien P, Polson N and Stephens D (Oxford University Press)

    2013

  34. [43]

    Nissinen A, Kolehmainen V and Kaipio J P 2011 International Journal for Uncertainty Quantification 1 203–222 ISSN 2152-5080

    2011

  35. [44]

    Koulouri A and Rimpil¨ ainen V 2020 Simultaneous skull conductivity and focal source imaging from EEG recordings with the help of bayesian uncertainty modelling 8th European Medical and Biological Engineering Conference (Springer International Publishing) pp 1019–1027

    2020

  36. [45]

    Lipponen A, Sepp¨ anen A and Kaipio J P 2011 Meas. Sci. Technol. 22 104013

    2011

  37. [46]

    Nissinen A, Heikkinen L M, Kolehmainen V and Kaipio J P 2009 Meas. Sci. Technol. 20 105504

    2009

  38. [47]

    Arridge S R, Kaipio J P, Kolehmainen V, Schweiger M, Somersalo E, Tarvainen T and Vauhkonen M 2006 Inverse Problems 22(1) 175–195

    2006

  39. [48]

    Kolehmainen V, Schweiger M, Nissil¨ a I, Tarvainen T, Arridge S R and Kaipio J P 2009 J. Opt. Soc. Am. A 26 2257–2268

    2009

  40. [49]

    Tarvainen T, Kolehmainen V, Pulkkinen A, Vauhkonen M, Schweiger M, Arridge S R and Kaipio J P 2010 Inverse Problems 26 015005

    2010

  41. [50]

    Koulouri A, Rimpil¨ ainen V, Brookes M and Kaipio J P 2016 Appl. Num. Math. 106 24–36

    2016

  42. [51]

    Mozumder M, Tarvainen T, Arridge S R, Kaipio J and Kolehmainen V 2013 Biomed. Opt. Express 4 2015–2031 URL http://www.opticsinfobase.org/boe/abstract.cfm?URI=boe-4-10-2015

    2013

  43. [52]

    org/10.1088/1361-6420/ad602e

    Alexanderian A, Nicholson R and Petra N 2024 Inverse Problems 40 095001 URL https://dx.doi. org/10.1088/1361-6420/ad602e

    2024 doi

  44. [53]

    Candiani V, Hyv¨ onen N, Kaipio J P and Kolehmainen V 2021 Inverse Problems 37 125008 URL https://dx.doi.org/10.1088/1361-6420/ac346a

    2021 doi

  45. [54]

    Nicholson R, Petra N, Villa U and Kaipio J P 2023 Inverse Problems 39 054001 URL https: //dx.doi.org/10.1088/1361-6420/acc129

    2023 doi

  46. [55]

    Bohrnstedt G W and Goldberger A S 1969 Journal of the American Statistical Association 64(328) 1439—-1442

    1969

  47. [56]

    Neuroeng

    Grech R, Cassar T, Muscat J, Camilleri K, Fabri S, Zervakis M, Xanthopoulos P, Sakkalis V and Vanrumste B 2008 J. Neuroeng. Rehabil. 5 25

    2008

  48. [57]

    Michel C M and Brunet D 2019 Frontiers in Neurology 10 ISSN 1664-2295 36

    2019

  49. [58]

    Fuchs M, Wagner M, Wischmann H A, K´’ohler T, Theissen A, Drenckhahn R and Buchner H 1998 Electroencephalogr. Clin. Neurophysiol. 107 93–111

    1998

  50. [59]

    Rodriguez-Rivera A, Van Veen B and Wakai R 2003IEEE Transactions on Biomedical Engineering 50 137–149

  51. [60]

    H¨ oltershinken M B, Erdbr¨ ugger T and Wolters C H 2024 (Preprint 2409.07459)

    2024 arXiv

  52. [61]

    Kn¨ osche T 1997The Netherlands

  53. [62]

    Vauhkonen M 1997 Electrical impedance tomography and prior information PhD thesis

    1997

  54. [63]

    Puonti O, Van Leemput K, Saturnino G B, Siebner H R, Madsen K H and Thielscher A 2020 NeuroImage 219 117044 ISSN 1053-8119 URL https://www.sciencedirect.com/science/article/ pii/S1053811920305309

    2020

  55. [64]

    Wolters C H, Grasedyck L and Hackbusch W 2004 Inverse Problems 20 1099–1116

    2004

  56. [65]

    16 29–38

    Hoekema R, Wieneke G H, van Veelen C W, van Rijen P C, Huiskamp G J, Ansems J and van Huffelen A C 2003 Brain Topogr. 16 29–38

    2003

  57. [66]

    Homma S, Musha T, Nakajima Y, Okomoto Y, Blom S, Flink R and Hagbarth K E 1995 Neurosci. Res. 22 51–55

    1995

  58. [67]

    Ramon C, Schimpf P H and Haueisen J 2006 Biomed. Eng. Online 5

    2006

  59. [68]

    Badia A E and Duong T H 1998 Inverse Problems 14 883–891 ISSN 1361-6420

    1998

  60. [69]

    K¨ ohler T, Wagner M, Fuchs M, Wischmann H A, Drenckhahn R and Theissen A 1996 Depth normalization in meg/eeg current density imaging Engineering in Medicine and Biology Society,

    1996

  61. [70]

    Pascual-Marqui R D, Michel C M and Lehmann D 1994 Int. J. Psychophysiol. 18 49–65

    1994

  62. [71]

    24, 5–12

    Pascual-Marqui R D 2002 Methods Find Exp Clin Pharmacol. 24, 5–12

    2002

  63. [72]

    Fuchs M, Wagner M and Wischmann H A 1999 Journal of Clinical Neurophysiology 16(3) 267–295

    1999

  64. [73]

    Wagner M, Wischmann H A, Fuchs M, K¨ ohler T and Drenckhahn R 2000 Current Density Reconstructions Using the L1 Norm (Springer New York) pp 393–396 ISBN 9781461212607

    2000

  65. [74]

    Palmero-Soler E, Dolan K, Hadamschek V and Tass P A 2007 Physics in Medicine and Biology 52 1783–1800 ISSN 1361-6560

    2007

  66. [75]

    Buchner H, Knoll G, Fuchs M, Rien ´’acker A, Beckmann R, Wagner M, Silny J and Pesch J 1997 Electroencephalogr. Clin. Neurophysiol. 102 267–278

    1997

  67. [76]

    Gramfort A, Luessi M, Larson E, Engemann D A, Strohmeier D, Brodbeck C, Parkkonen L and H¨ am¨ al¨ ainen M S 2014NeuroImage 86 446–460 ISSN 1053-8119

  68. [77]

    Koulouri A, Rimpil¨ ainen V, Brookes M and Kaipio J P 2018 Prior variances and depth un- biased estimators in eeg focal source imaging EMBEC & NBC 2017 (Springer) pp 33–36 URL https://doi.org/10.1007/978-981-10-5122-7 9

    2018 doi

  69. [78]

    Calvetti D, Pascarella A, Pitolli F, Somersalo E and Vantaggi B 2018Brain Topography32 363–393 ISSN 1573-6792

  70. [1996]

    Proceedings of the 18th Annual International Conference of the IEEE vol 2 pp 812–813 vol.2

    Bridging Disciplines for Biomedicine. Proceedings of the 18th Annual International Conference of the IEEE vol 2 pp 812–813 vol.2

Pith tools