Neurotechnology · Computational neuroscience
My work focuses on neural signal acquisition, real-time processing and decoding, with applications in medical technology and brain–computer interfaces.
It runs from methods research on real-time source imaging, through open-source scientific software, to the development of medical systems in industry.
Head of Physical AI & Computing, ZEISS
Auditory evoked response. Source estimate (dSPM) of magnetoencephalography recorded while tones were played to the right ear, shown on the cortical surface of the participant, left hemisphere. The trace is the root-mean-square signal of the planar gradiometers, peaking at 42 fT/cm.
dSPM 5–18MNE sample dataset, average of 61 trials. Computed with MNE-Python.
Three connected areas. The first is established work with a published record, the third continues it, and the second is the direction I am moving towards.
01Established work
Magneto- and electroencephalography follow brain activity with millisecond resolution, but locating its sources is an ill-posed inverse problem that is normally solved after the recording. My doctoral and postdoctoral work made source estimates available during the measurement: a reduction of the source space based on a cortical atlas, GPU implementations of scanning methods, and the acquisition software needed to run them on clinical systems.
02Research interest
Interfaces that stimulate as well as record have to do so selectively and without corrupting the recording. Combining electrical read-out with optical methods is one candidate, and an open engineering problem: light delivery, heating, photoelectric artefacts and stability over years all have to be controlled. I am interested in how such interfaces can be modelled and evaluated before they are built. This is planned work; I report no results here.
03Current research
Decoding turns measured activity into something usable: a state, an intention, a control signal. My earlier work used recurrent networks to bring temporal context into source estimates. Current work concerns decoding from non-invasive recordings and the real-time systems needed to close the loop between measurement and response.
Since 2010 · Open-source software
An open-source C++ framework for real-time and offline processing of MEG, EEG and related electrophysiological data: libraries for file handling, forward and inverse modelling and signal processing, and applications for acquisition, analysis and 3D visualisation. I initiated the project and co-lead its development. More than forty people have contributed to it.
ProjectRepository (opens in a new tab)Run in the browser (opens in a new tab)
2017–2025
Acquisition software for magnetic resonance imaging: a language for pulse sequences, a receiver implemented on a graphics card, and an open platform for scanner control.
Since 2026
An independent editorial project that follows research, clinical studies and companies in brain–computer interfaces, with a source for every claim.
2015–today
Surgical navigation, neonatal MRI, cardiac mapping and platforms for machine learning in regulated products, as architect and team lead.
2021
Dinh C, Samuelsson JG, Hunold A, Hämäläinen MS, Khan S. Contextual MEG and EEG source estimates using spatiotemporal LSTM networks. Frontiers in Neuroscience 2021;15:552666.
doi:10.3389/fnins.2021.552666 (opens in a new tab)Open accessCode (opens in a new tab)
2019
Esch L, Dinh C, Larson E, Engemann D, Jas M, Khan S, Gramfort A, Hämäläinen MS. MNE: Software for acquiring, processing, and visualizing MEG/EEG data. Magnetoencephalography: From Signals to Dynamic Cortical Networks, 2nd ed. (Supek S, Aine CJ, eds.), Springer 2019, pp. 355–371.
doi:10.1007/978-3-030-00087-5_59 (opens in a new tab)Author manuscript (HAL) (opens in a new tab)
2018
Esch L, Sun L, Klüber V, Lew S, Baumgarten D, Grant PE, Okada Y, Haueisen J, Hämäläinen MS, Dinh C. MNE Scan: Software for real-time processing of electrophysiological data. Journal of Neuroscience Methods 2018;303:55–67.
doi:10.1016/j.jneumeth.2018.03.020 (opens in a new tab)Author manuscript (PMC) (opens in a new tab)Code (opens in a new tab)
2016
Okada Y, Hämäläinen M, Pratt K, Mascarenas A, Miller P, Han M, Robles J, Cavallini A, Power B, Sieng K, Sun L, Lew S, Doshi C, Ahtam B, Dinh C, Esch L, Grant E, Nummenmaa A, Paulson D. BabyMEG: A whole-head pediatric magnetoencephalography system for human brain development research. Review of Scientific Instruments 2016;87(9):094301.
2015
Dinh C, Strohmeier D, Luessi M, Güllmar D, Baumgarten D, Haueisen J, Hämäläinen MS. Real-time MEG source localization using regional clustering. Brain Topography 2015;28(6):771–784.
doi:10.1007/s10548-015-0431-9 (opens in a new tab)Author manuscript (PMC) (opens in a new tab)
I am a biomedical engineer. I studied at Technische Universität Ilmenau and completed a doctorate there in 2015 on the real-time localisation of neuronal activity, with research stays and later a postdoctoral fellowship at the Athinoula A. Martinos Center for Biomedical Imaging at Massachusetts General Hospital and Harvard Medical School.
Since then I have developed medical systems in industry: surgical navigation at Stryker, magnetic resonance imaging at Neoscan Solutions and cardiac mapping at Ablacon. I now lead Physical AI & Computing at the ZEISS Innovation Hub @ KIT (opens in a new tab) in Karlsruhe, where I also coordinate the company’s technology field of brain–computer interfaces.