Mihyun Choi

All Publications

• Optimized ultrasound neuromodulation for non-invasive control of behavior and physiology. Neuron Murphy, K. R., Farrell, J. S., Bendig, J., Mitra, A., Luff, C., Stelzer, I. A., Yamaguchi, H., Angelakos, C. C., Choi, M., Bian, W., DiIanni, T., Pujol, E. M., Matosevich, N., Airan, R., Gaudillière, B., Konofagou, E. E., Butts-Pauly, K., Soltesz, I., de Lecea, L. 2024

  Abstract

  Focused ultrasound can non-invasively modulate neural activity, but whether effective stimulation parameters generalize across brain regions and cell types remains unknown. We used focused ultrasound coupled with fiber photometry to identify optimal neuromodulation parameters for four different arousal centers of the brain in an effort to yield overt changes in behavior. Applying coordinate descent, we found that optimal parameters for excitation or inhibition are highly distinct, the effects of which are generally conserved across brain regions and cell types. Optimized stimulations induced clear, target-specific behavioral effects, whereas non-optimized protocols of equivalent energy resulted in substantially less or no change in behavior. These outcomes were independent of auditory confounds and, contrary to expectation, accompanied by a cyclooxygenase-dependent and prolonged reduction in local blood flow and temperature with brain-region-specific scaling. These findings demonstrate that carefully tuned and targeted ultrasound can exhibit powerful effects on complex behavior and physiology.

  View details for DOI 10.1016/j.neuron.2024.07.002

  View details for PubMedID 39079529

• Development and validation of a computational method to predict unintended auditory brainstem response during transcranial ultrasound neuromodulation in mice. Brain stimulation Choi, M. H., Li, N., Popelka, G., Butts Pauly, K. 2023

  Abstract

  Transcranial ultrasound stimulation (TUS) is a promising noninvasive neuromodulation modality. The inadvertent and unpredictable activation of the auditory system in response to TUS obfuscates the interpretation of non-auditory neuromodulatory responses.The objective was to develop and validate a computational metric to quantify the susceptibility to unintended auditory brainstem response (ABR) in mice premised on time frequency analyses of TUS signals and auditory sensitivity.Ultrasound pulses with varying amplitudes, pulse repetition frequencies (PRFs), envelope smoothing profiles, and sinusoidal modulation frequencies were selected. Each pulse's time-varying frequency spectrum was differentiated across time, weighted by the mouse hearing sensitivity, then summed across frequencies. The resulting time-varying function, computationally predicting the ABR, was validated against experimental ABR in mice during TUS with the corresponding pulse.There was a significant correlation between experimental ABRs and the computational predictions for 19 TUS signals (R2 = 0.97).To reduce ABR in mice during in vivo TUS studies, 1) reduce the amplitude of a rectangular continuous wave envelope, 2) increase the rise/fall times of a smoothed continuous wave envelope, and/or 3) change the PRF and/or duty cycle of a rectangular or sinusoidal pulsed wave to reduce the gap between pulses and increase the rise/fall time of the overall envelope. This metric can aid researchers performing in vivo mouse studies in selecting TUS signal parameters that minimize unintended ABR. The methods for developing this metric can be adapted to other animal models.

  View details for DOI 10.1016/j.brs.2023.09.004

  View details for PubMedID 37690602

• Measuring sound velocity based on acoustic resonance using multiple narrow band transducers. Heliyon Li, N., Chen, Z., Zhu, J., Choi, M. H., Yang, J., Yuan, Z., Sun, L., Fei, C., Qiu, Z. 2023; 9 (3): e14227

  Abstract

  The sound velocity in a medium is closely related to its material properties, including its composition, structure, density, pressure, and temperature. Various methods have been developed to determine the sound velocity through materials. Among them, a strategy based on ultrasound resonance frequency has been most widely used due to the simplicity. However, it requires a transducer with a wide bandwidth to cover enough resonance frequencies to perform the consequent calculations. In this paper, we develop a resonance method for measuring sound velocity, using multi-frequency narrow-band transducers breaking through the limitation of transducer bandwidth on the utilization of the resonance method. We use different transducers at different center frequencies and with different bandwidth to measure the sound velocity in 100-mum and 400-mum thick steel pieces. The measurement results of different combinations are in good agreement, verifying that the use of multi-frequency narrow-band transducer combinations. Given that most therapeutic transducers have a narrow bandwidth, this method can be used during intracranial ultrasound stimulation to optimize targeting by non-invasively measuring the sound velocity in the skull, especially at thinner locations.

  View details for DOI 10.1016/j.heliyon.2023.e14227

  View details for PubMedID 36950590

• Sonogenetics: Recent advances and future directions. Brain stimulation Liu, T., Choi, M. H., Zhu, J., Zhu, T., Yang, J., Li, N., Chen, Z., Xian, Q., Hou, X., He, D., Guo, J., Fei, C., Sun, L., Qiu, Z. 2022

  Abstract

  Sonogenetics refers to the use of genetically encoded, ultrasound-responsive mediators for noninvasive and selective control of neural activity. It is a promising tool for studying neural circuits. However, due to its infancy, basic studies and developments are still underway, including gauging key in vivo performance metrics such as spatiotemporal resolution, selectivity, specificity, and safety. In this paper, we summarize recent findings on sonogenetics to highlight technical hurdles that have been cleared, challenges that remain, and future directions for optimization.

  View details for DOI 10.1016/j.brs.2022.09.002

  View details for PubMedID 36130679

• Predicting malnutrition from longitudinal patient trajectories with deep learning. PloS one Jin, B. T., Choi, M. H., Moyer, M. F., Kim, D. A. 2022; 17 (7): e0271487

  Abstract

  Malnutrition is common, morbid, and often correctable, but subject to missed and delayed diagnosis. Better screening and prediction could improve clinical, functional, and economic outcomes. This study aimed to assess the predictability of malnutrition from longitudinal patient records, and the external generalizability of a predictive model. Predictive models were developed and validated on statewide emergency department (ED) and hospital admission databases for California, Florida and New York, including visits from October 1, 2015 to December 31, 2018. Visit features included patient demographics, diagnosis codes, and procedure categories. Models included long short-term memory (LSTM) recurrent neural networks trained on longitudinal trajectories, and gradient-boosted tree and logistic regression models trained on cross-sectional patient data. The dataset used for model training and internal validation (California and Florida) included 62,811 patient trajectories (266,951 visits). Test sets included 63,997 (California), 63,112 (Florida), and 62,472 (New York) trajectories, such that each cohort's composition was proportional to the prevalence of malnutrition in that state. Trajectories contained seven patient characteristics and up to 2,008 diagnosis categories. Area under the receiver-operating characteristic (AUROC) and precision-recall curves (AUPRC) were used to characterize prediction of first malnutrition diagnoses in the test sets. Data analysis was performed from September 2020 to May 2021. Between 4.0% (New York) and 6.2% (California) of patients received malnutrition diagnoses. The longitudinal LSTM model produced the most accurate predictions of malnutrition, with comparable predictive performance in California (AUROC 0.854, AUPRC 0.258), Florida (AUROC 0.869, AUPRC 0.234), and New York (AUROC 0.869, AUPRC 0.190). Deep learning models can reliably predict malnutrition from existing longitudinal patient records, with better predictive performance and lower data-collection requirements than existing instruments. This approach may facilitate early nutritional intervention via automated screening at the point of care.

  View details for DOI 10.1371/journal.pone.0271487

  View details for PubMedID 35901027