SEMPAI: a Self‐Enhancing Multi‐Photon Artificial Intelligence for Prior‐Informed Assessment of Muscle Function and Pathology
Abstract Deep learning (DL) shows notable success in biomedical studies. However, most DL algorithms work as black boxes, exclude biomedical experts, and need extensive data. This is especially problematic for fundamental research in the laboratory, where often only small and sparse data are availab...
| Main Authors: | , , , , , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2023-10-01
|
| Series: | Advanced Science |
| Subjects: | |
| Online Access: | https://doi.org/10.1002/advs.202206319 |
| _version_ | 1797663860961837056 |
|---|---|
| author | Alexander Mühlberg Paul Ritter Simon Langer Chloë Goossens Stefanie Nübler Dominik Schneidereit Oliver Taubmann Felix Denzinger Dominik Nörenberg Michael Haug Sebastian Schürmann Roarke Horstmeyer Andreas K. Maier Wolfgang H. Goldmann Oliver Friedrich Lucas Kreiss |
| author_facet | Alexander Mühlberg Paul Ritter Simon Langer Chloë Goossens Stefanie Nübler Dominik Schneidereit Oliver Taubmann Felix Denzinger Dominik Nörenberg Michael Haug Sebastian Schürmann Roarke Horstmeyer Andreas K. Maier Wolfgang H. Goldmann Oliver Friedrich Lucas Kreiss |
| author_sort | Alexander Mühlberg |
| collection | DOAJ |
| description | Abstract Deep learning (DL) shows notable success in biomedical studies. However, most DL algorithms work as black boxes, exclude biomedical experts, and need extensive data. This is especially problematic for fundamental research in the laboratory, where often only small and sparse data are available and the objective is knowledge discovery rather than automation. Furthermore, basic research is usually hypothesis‐driven and extensive prior knowledge (priors) exists. To address this, the Self‐Enhancing Multi‐Photon Artificial Intelligence (SEMPAI) that is designed for multiphoton microscopy (MPM)‐based laboratory research is presented. It utilizes meta‐learning to optimize prior (and hypothesis) integration, data representation, and neural network architecture simultaneously. By this, the method allows hypothesis testing with DL and provides interpretable feedback about the origin of biological information in 3D images. SEMPAI performs multi‐task learning of several related tasks to enable prediction for small datasets. SEMPAI is applied on an extensive MPM database of single muscle fibers from a decade of experiments, resulting in the largest joint analysis of pathologies and function for single muscle fibers to date. It outperforms state‐of‐the‐art biomarkers in six of seven prediction tasks, including those with scarce data. SEMPAI's DL models with integrated priors are superior to those without priors and to prior‐only approaches. |
| first_indexed | 2024-03-11T19:21:59Z |
| format | Article |
| id | doaj.art-44bff05520d54a82863b2dbf105f2848 |
| institution | Directory Open Access Journal |
| issn | 2198-3844 |
| language | English |
| last_indexed | 2024-03-11T19:21:59Z |
| publishDate | 2023-10-01 |
| publisher | Wiley |
| record_format | Article |
| series | Advanced Science |
| spelling | doaj.art-44bff05520d54a82863b2dbf105f28482023-10-07T03:51:50ZengWileyAdvanced Science2198-38442023-10-011028n/an/a10.1002/advs.202206319SEMPAI: a Self‐Enhancing Multi‐Photon Artificial Intelligence for Prior‐Informed Assessment of Muscle Function and PathologyAlexander Mühlberg0Paul Ritter1Simon Langer2Chloë Goossens3Stefanie Nübler4Dominik Schneidereit5Oliver Taubmann6Felix Denzinger7Dominik Nörenberg8Michael Haug9Sebastian Schürmann10Roarke Horstmeyer11Andreas K. Maier12Wolfgang H. Goldmann13Oliver Friedrich14Lucas Kreiss15Institute of Medical Biotechnology Department of Chemical and Biological Engineering Friedrich‐Alexander University Erlangen‐Nuremberg Paul‐Gordan‐Str. 3 91052 Erlangen GermanyInstitute of Medical Biotechnology Department of Chemical and Biological Engineering Friedrich‐Alexander University Erlangen‐Nuremberg Paul‐Gordan‐Str. 3 91052 Erlangen GermanyPattern Recognition Lab Department of Computer Science Friedrich‐Alexander University Erlangen‐Nuremberg Martensstr. 3 91058 Erlangen GermanyClinical Division and Laboratory of Intensive Care Medicine KU Leuven UZ Herestraat 49 – P.O. box 7003 Leuven 3000 BelgiumInstitute of Medical Biotechnology Department of Chemical and Biological Engineering Friedrich‐Alexander University Erlangen‐Nuremberg Paul‐Gordan‐Str. 3 91052 Erlangen GermanyInstitute of Medical Biotechnology Department of Chemical and Biological Engineering Friedrich‐Alexander University Erlangen‐Nuremberg Paul‐Gordan‐Str. 3 91052 Erlangen GermanyPattern Recognition Lab Department of Computer Science Friedrich‐Alexander University Erlangen‐Nuremberg Martensstr. 3 91058 Erlangen GermanyPattern Recognition Lab Department of Computer Science Friedrich‐Alexander University Erlangen‐Nuremberg Martensstr. 3 91058 Erlangen GermanyDepartment of Radiology and Nuclear Medicine University Medical Center Mannheim Medical Faculty Mannheim Theodor‐Kutzer‐Ufer 1–3 68167 Mannheim GermanyInstitute of Medical Biotechnology Department of Chemical and Biological Engineering Friedrich‐Alexander University Erlangen‐Nuremberg Paul‐Gordan‐Str. 3 91052 Erlangen GermanyInstitute of Medical Biotechnology Department of Chemical and Biological Engineering Friedrich‐Alexander University Erlangen‐Nuremberg Paul‐Gordan‐Str. 3 91052 Erlangen GermanyComputational Optics Lab Department of Biomedical Engineering Duke University 101 Science Dr Durham NC 27708 USAPattern Recognition Lab Department of Computer Science Friedrich‐Alexander University Erlangen‐Nuremberg Martensstr. 3 91058 Erlangen GermanyBiophysics Group Department of Physics Friedrich‐Alexander University Erlangen‐Nuremberg Henkestr. 91 91052 Erlangen GermanyInstitute of Medical Biotechnology Department of Chemical and Biological Engineering Friedrich‐Alexander University Erlangen‐Nuremberg Paul‐Gordan‐Str. 3 91052 Erlangen GermanyInstitute of Medical Biotechnology Department of Chemical and Biological Engineering Friedrich‐Alexander University Erlangen‐Nuremberg Paul‐Gordan‐Str. 3 91052 Erlangen GermanyAbstract Deep learning (DL) shows notable success in biomedical studies. However, most DL algorithms work as black boxes, exclude biomedical experts, and need extensive data. This is especially problematic for fundamental research in the laboratory, where often only small and sparse data are available and the objective is knowledge discovery rather than automation. Furthermore, basic research is usually hypothesis‐driven and extensive prior knowledge (priors) exists. To address this, the Self‐Enhancing Multi‐Photon Artificial Intelligence (SEMPAI) that is designed for multiphoton microscopy (MPM)‐based laboratory research is presented. It utilizes meta‐learning to optimize prior (and hypothesis) integration, data representation, and neural network architecture simultaneously. By this, the method allows hypothesis testing with DL and provides interpretable feedback about the origin of biological information in 3D images. SEMPAI performs multi‐task learning of several related tasks to enable prediction for small datasets. SEMPAI is applied on an extensive MPM database of single muscle fibers from a decade of experiments, resulting in the largest joint analysis of pathologies and function for single muscle fibers to date. It outperforms state‐of‐the‐art biomarkers in six of seven prediction tasks, including those with scarce data. SEMPAI's DL models with integrated priors are superior to those without priors and to prior‐only approaches.https://doi.org/10.1002/advs.202206319deep learningexplainable artificial intelligencemeta‐learningmultiphoton microscopymuscle researchprior information integration |
| spellingShingle | Alexander Mühlberg Paul Ritter Simon Langer Chloë Goossens Stefanie Nübler Dominik Schneidereit Oliver Taubmann Felix Denzinger Dominik Nörenberg Michael Haug Sebastian Schürmann Roarke Horstmeyer Andreas K. Maier Wolfgang H. Goldmann Oliver Friedrich Lucas Kreiss SEMPAI: a Self‐Enhancing Multi‐Photon Artificial Intelligence for Prior‐Informed Assessment of Muscle Function and Pathology Advanced Science deep learning explainable artificial intelligence meta‐learning multiphoton microscopy muscle research prior information integration |
| title | SEMPAI: a Self‐Enhancing Multi‐Photon Artificial Intelligence for Prior‐Informed Assessment of Muscle Function and Pathology |
| title_full | SEMPAI: a Self‐Enhancing Multi‐Photon Artificial Intelligence for Prior‐Informed Assessment of Muscle Function and Pathology |
| title_fullStr | SEMPAI: a Self‐Enhancing Multi‐Photon Artificial Intelligence for Prior‐Informed Assessment of Muscle Function and Pathology |
| title_full_unstemmed | SEMPAI: a Self‐Enhancing Multi‐Photon Artificial Intelligence for Prior‐Informed Assessment of Muscle Function and Pathology |
| title_short | SEMPAI: a Self‐Enhancing Multi‐Photon Artificial Intelligence for Prior‐Informed Assessment of Muscle Function and Pathology |
| title_sort | sempai a self enhancing multi photon artificial intelligence for prior informed assessment of muscle function and pathology |
| topic | deep learning explainable artificial intelligence meta‐learning multiphoton microscopy muscle research prior information integration |
| url | https://doi.org/10.1002/advs.202206319 |
| work_keys_str_mv | AT alexandermuhlberg sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT paulritter sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT simonlanger sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT chloegoossens sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT stefanienubler sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT dominikschneidereit sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT olivertaubmann sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT felixdenzinger sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT dominiknorenberg sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT michaelhaug sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT sebastianschurmann sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT roarkehorstmeyer sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT andreaskmaier sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT wolfganghgoldmann sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT oliverfriedrich sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology AT lucaskreiss sempaiaselfenhancingmultiphotonartificialintelligenceforpriorinformedassessmentofmusclefunctionandpathology |