
Yuanyuan Gao
Postdoctoral Scholar, Psychiatry
Bio
Yuanyuan Gao completed her PhD at Rensselaer Polytechnic Institute. Her PhD thesis researched the effects of neuromodulation on human motor learning using functional near-infrared spectroscopy (fNIRS). She finished her first postdoctoral training term in Dr. David Boas' lab in Boston University on advanced fNIRS data analysis. She is now a postdoctoral fellow working at Stanford University for her second term of postdoctoral training on the clinical applications of fNIRS. Her research interests are fNIRS, its multimodels with fMRI, EEG, eye-tracker, physiology measurements, neuromodulation and machine learning models, and its applications in clinical research.
Professional Education
-
Master of Engineering, High School Affiliated to Beihang University (2013)
-
Doctor of Philosophy, Rensselaer Polytechnic Institute (2020)
-
Bachelor of Engineering, High School Affiliated to Beihang University (2010)
-
Ph.D., Rensselaer Polytechnic Institute, Mechanical Engineering (2020)
-
Master, Beihang University, Mechanical Engineering (2013)
-
BS, Beihang University, Aircraft Environment and Life Security Engineering (2010)
All Publications
-
Deep learning-based motion artifact removal in functional near-infrared spectroscopy
NEUROPHOTONICS
2022; 9 (4): 041406
Abstract
Significance: Functional near-infrared spectroscopy (fNIRS), a well-established neuroimaging technique, enables monitoring cortical activation while subjects are unconstrained. However, motion artifact is a common type of noise that can hamper the interpretation of fNIRS data. Current methods that have been proposed to mitigate motion artifacts in fNIRS data are still dependent on expert-based knowledge and the post hoc tuning of parameters. Aim: Here, we report a deep learning method that aims at motion artifact removal from fNIRS data while being assumption free. To the best of our knowledge, this is the first investigation to report on the use of a denoising autoencoder (DAE) architecture for motion artifact removal. Approach: To facilitate the training of this deep learning architecture, we (i) designed a specific loss function and (ii) generated data to mimic the properties of recorded fNIRS sequences. Results: The DAE model outperformed conventional methods in lowering residual motion artifacts, decreasing mean squared error, and increasing computational efficiency. Conclusion: Overall, this work demonstrates the potential of deep learning models for accurate and fast motion artifact removal in fNIRS data.
View details for DOI 10.1117/1.NPh.9.4.041406
View details for Web of Science ID 000926179900006
View details for PubMedID 35475257
View details for PubMedCentralID PMC9034734
-
Functional Brain Imaging Reliably Predicts Bimanual Motor Skill Performance in a Standardized Surgical Task
IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING
2021; 68 (7): 2058-2066
Abstract
Currently, there is a dearth of objective metrics for assessing bi-manual motor skills, which are critical for high-stakes professions such as surgery. Recently, functional near-infrared spectroscopy (fNIRS) has been shown to be effective at classifying motor task types, which can be potentially used for assessing motor performance level. In this work, we use fNIRS data for predicting the performance scores in a standardized bi-manual motor task used in surgical certification and propose a deep-learning framework 'Brain-NET' to extract features from the fNIRS data. Our results demonstrate that the Brain-NET is able to predict bi-manual surgical motor skills based on neuroimaging data accurately ( R2=0.73). Furthermore, the classification ability of the Brain-NET model is demonstrated based on receiver operating characteristic (ROC) curves and area under the curve (AUC) values of 0.91. Hence, these results establish that fNIRS associated with deep learning analysis is a promising method for a bedside, quick and cost-effective assessment of bi-manual skill levels.
View details for DOI 10.1109/TBME.2020.3014299
View details for Web of Science ID 000663531500002
View details for PubMedID 32755850
View details for PubMedCentralID PMC8265734
-
Decreasing the Surgical Errors by Neurostimulation of Primary Motor Cortex and the Associated Brain Activation via Neuroimaging
FRONTIERS IN NEUROSCIENCE
2021; 15: 651192
Abstract
Acquisition of fine motor skills is a time-consuming process as it is based on learning via frequent repetitions. Transcranial electrical stimulation (tES) is a promising means of enhancing simple motor skill development via neuromodulatory mechanisms. Here, we report that non-invasive neurostimulation facilitates the learning of complex fine bimanual motor skills associated with a surgical task. During the training of 12 medical students on the Fundamentals of Laparoscopic Surgery (FLS) pattern cutting task over a period of 12 days, we observed that transcranial direct current stimulation (tDCS) decreased error level and the variability in performance, compared to the Sham group. Furthermore, by concurrently monitoring the cortical activations of the subjects via functional near-infrared spectroscopy (fNIRS), our study showed that the cortical activation patterns were significantly different between the tDCS and Sham group, with the activation of primary motor cortex (M1) and prefrontal cortex (PFC) contralateral to the anodal electrode significantly decreased while supplemental motor area (SMA) increased by tDCS. The lowered performance errors were retained after 1-month post-training. This work supports the use of tDCS to enhance performance accuracy in fine bimanual motor tasks.
View details for DOI 10.3389/fnins.2021.651192
View details for Web of Science ID 000636667300001
View details for PubMedID 33828456
View details for PubMedCentralID PMC8019915
-
Functional brain connectivity related to surgical skill dexterity in physical and virtual simulation environments
NEUROPHOTONICS
2021; 8 (1): 015008
Abstract
Significance: Surgical simulators, both virtual and physical, are increasingly used as training tools for teaching and assessing surgical technical skills. However, the metrics used for assessment in these simulation environments are often subjective and inconsistent. Aim: We propose functional activation metrics, derived from brain imaging measurements, to objectively assess the correspondence between brain activation with surgical motor skills for subjects with varying degrees of surgical skill. Approach: Cortical activation based on changes in the oxygenated hemoglobin (HbO) of 36 subjects was measured using functional near-infrared spectroscopy at the prefrontal cortex (PFC), primary motor cortex, and supplementary motor area (SMA) due to their association with motor skill learning. Inter-regional functional connectivity metrics, namely, wavelet coherence (WCO) and wavelet phase coherence were derived from HbO changes to correlate brain activity to surgical motor skill levels objectively. Results: One-way multivariate analysis of variance found a statistically significant difference in the inter-regional WCO metrics for physical simulator based on Wilk's Λ for expert versus novice, F ( 10,1 ) = 7495.5 , p < 0.01 . Partial eta squared effect size for the inter-regional WCO metrics was found to be highest between the central prefrontal cortex (CPFC) and SMA, CPFC-SMA ( η 2 = 0.257 ). Two-tailed Mann-Whitney U tests with a 95% confidence interval showed baseline equivalence and a statistically significant ( p < 0.001 ) difference in the CPFC-SMA WPCO metrics for the physical simulator training group ( 0.960 ± 0.045 ) versus the untrained control group ( 0.735 ± 0.177 ) following training for 10 consecutive days in addition to the pretest and posttest days. Conclusion: We show that brain functional connectivity WCO metric corresponds to surgical motor skills in the laparoscopic physical simulators. Functional connectivity between the CPFC and the SMA is lower for subjects that exhibit expert surgical motor skills than untrained subjects in laparoscopic physical simulators.
View details for DOI 10.1117/1.NPh.8.1.015008
View details for Web of Science ID 000636677100012
View details for PubMedID 33681406
View details for PubMedCentralID PMC7927423
-
The Effects of Transcranial Electrical Stimulation on Human Motor Functions: A Comprehensive Review of Functional Neuroimaging Studies
FRONTIERS IN NEUROSCIENCE
2020; 14: 744
Abstract
Transcranial electrical stimulation (tES) is a promising tool to enhance human motor skills. However, the underlying physiological mechanisms are not fully understood. On the other hand, neuroimaging modalities provide powerful tools to map some of the neurophysiological biomarkers associated with tES. Here, a comprehensive review was undertaken to summarize the neuroimaging evidence of how tES affects human motor skills. A literature search has been done on the PubMed database, and 46 relative articles were selected. After reviewing these articles, we conclude that neuroimaging techniques are feasible to be coupled with tES and offer valuable information of cortical excitability, connectivity, and oscillations regarding the effects of tES on human motor behavior. The biomarkers derived from neuroimaging could also indicate the motor performance under tES conditions. This approach could advance the understanding of tES effects on motor skill and shed light on a new generation of adaptive stimulation models.
View details for DOI 10.3389/fnins.2020.00744
View details for Web of Science ID 000560717500001
View details for PubMedID 32792898
View details for PubMedCentralID PMC7393222
-
A machine learning approach to predict surgical learning curves
MOSBY-ELSEVIER. 2020: 321-327
Abstract
Contemporary surgical training programs rely on the repetition of selected surgical motor tasks. Such methodology is inherently open ended with no control on the time taken to attain a set level of proficiency, given the trainees' intrinsic differences in initial skill levels and learning abilities. Hence, an efficient training program should aim at tailoring the surgical training protocols to each trainee. In this regard, a predictive model using information from the initial learning stage to predict learning curve characteristics should facilitate the whole surgical training process.This paper analyzes learning curve data to train a multivariate supervised machine learning model. One factor is extracted to define the trainees' learning ability. An unsupervised machine learning model is also utilized for trainee classification. When established, the model can predict robustly the learning curve characteristics based on the first few trials.We show that the information present in the first 10 trials of surgical tasks can be utilized to predict the number of trials required to achieve proficiency (R2=0.72) and the final performance level (R2=0.89). Furthermore, only a single factor, learning index, is required to describe the learning process and to classify learners with unique learning characteristics.Using machine learning models, we show, for the first time, that the first few trials contain sufficient information to predict learning curve characteristics and that a single factor can capture the complex learning behavior. Using such models holds the potential for personalization of training regimens, leading to greater efficiency and lower costs.
View details for DOI 10.1016/j.surg.2019.10.008
View details for Web of Science ID 000510532300017
View details for PubMedID 31753325
View details for PubMedCentralID PMC6980926