LFP signals and spectrums: Signal to interference ratio (SIR): 𝑺𝒑𝒆𝒄𝒕𝒓𝒖𝒎 𝒂𝒎𝒑𝒍𝒊𝒕𝒖𝒅𝒆 𝒐𝒇 𝑳𝑭𝑷 𝒂𝒕 𝒖𝒍𝒕𝒓𝒂𝒔𝒐𝒖𝒏𝒅 𝒄𝒆𝒏𝒕𝒆𝒓 𝒇𝒓𝒆𝒒𝒖𝒆𝒏𝒄𝒚 𝑰𝒏𝒕𝒆𝒓𝒇𝒆𝒓𝒆𝒏𝒄𝒆 𝒍𝒆𝒗𝒆𝒍 𝒂𝒕 𝒖𝒍𝒕𝒓𝒂𝒔𝒐𝒖𝒏𝒅 𝒄𝒆𝒏𝒕𝒆𝒓 𝒇𝒓𝒆𝒒𝒖𝒆𝒏𝒄𝒚 •Lower ultrasound frequency ➔Higher displacement for a given pressure •Higher displacement ➔higher SIR •Lower ultrasound frequency ➔higher interference Ultrasound frequency optimization for highest SIR Acousto-electric Frequency Shifting and Filtering of Deep Neuronal Activity: A New Technique for Acousto-Electrophysiological Neuroimaging Mehdi Soozande1, Emmeric Tanghe1, Thomas Tarnaud1, 2 1 WAVES, Department of Information Technology (INTEC), Ghent University/IMEC, Tech Lane 126, 9052 Ghent, Belgium 2 4BRAIN, Department of Neurology, Ghent University Hospital, Corneel Heymanslaan 10, 9000 Ghent, Belgium Introduction Electrophysiological neuroimaging is critical for understanding brain function and diagnosing neurological disorders. Challenge: •EEG/MEG → high temporal resolution, poor spatial specificity •fMRI → better spatial resolution, but slow and indirect Goal: Non-invasive recording of deep brain activity with high spatial and temporal resolution Solution –Acousto-electrophysiological Neuroimaging (AENI): •Combines focused ultrasound (for deep, localized targeting) with electrophysiology (for direct neuronal signals) •Ultrasound induces microscopic vibrations of neurons in the target region •Vibrations cause frequency shifting of neural signals into sidebands around the ultrasound carrier •Signals from target region → shifted and detectable •Signals from non-target regions → remain unshifted, can be suppressed Computational modelling: •Validate the physical feasibility of ultrasound-induced frequency shifting. •Optimize acoustic and electrophysiological parameters safely and efficiently. •Predict signal characteristics and artifacts under controlled conditions. Impact: •Enables selective extraction of deep neuronal activity •Enhances spatial resolution while maintaining millisecond temporal resolution •Potential applications in neuroscience research and clinical diagnostics Generating Acousto-Electric LFP Signals Neuron Model: •Implemented a simple ball-and-stick neuron in NetPyNE •Stimulated electrically via somatic current injection •Electrode Placement: Virtual LFP electrode positioned 50 µm from the middle of the axon (ref. electrode: infinity) Ultrasound Modelling: •Planar ultrasound wave applied as a homogeneous vibration of the neuron relative to the electrode •Displacement amplitude determined by applied ultrasound pressure •Vibration frequency matched to the ultrasound carrier frequency Signal Generation: vibration of neuron changes electrode-neuron distance, producing acousto-electric frequency shifting of recorded LFP signals Decoding Targeted Neural Activity Bandpass filter •IIR 20th-order Butterworth filter •Center frequency: ultrasound frequency (0.5 MHz, 1 MHz) •Bandwidth: 30 kHz Lowpass filter •IIR Butterworth filter •Passband: 1 kHz •Stopband: 2 kHz •Stopband attenuation: 60 dB Demodulation •Frequency: ultrasound central frequency Contact
[email protected] Acknowledgement This project has received funding from the ERC Starting Grant (URENIMOD, 101162708). Funded by the European Union. Views and opinions expressed are however those of the authors only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them. BPF sin(2𝜋𝑓 𝐴𝑡) ×LPF Recorded signal Recovered signal Figure 2: LFP recording of a single neuron (Created by ChatGPT) Figure 3: ultrasound-induced neuron vibration and LFP recording (Created by ChatGPT) Figure 1: AENI: integration of focused ultrasound with electrophysiological recording (Created by Biorender ) INTRODUCTION RESULTS CONCLUSIONS FUTURE WORKS METHODS •More realistic simulations: morphologically realistic cells, realistic pressure field distributions •optimization of protocols and electrodes/transducers to maximize SNR and SIR •In vivo and in vitro experimental validation •AENI leverages ultrasound-induced frequency shifting for spatially selective neural recording. •Frequency-domain decoding isolates activity from targeted regions while suppressing interference. •NetPyNE simulations show up to 25 dB SIR improvement at subthreshold ultrasound pressures.
