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.
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)
Allan Reiss, Postdoctoral Faculty Sponsor
Prenatal and childhood exposure to organophosphate pesticides and functional brain imaging in young adults.
Early life exposure to organophosphate (OP) pesticides has been linked with poorer neurodevelopment from infancy to adolescence. In our Center for the Health Assessment of Mothers and Children of Salinas (CHAMACOS) birth cohort, we previously reported that residential proximity to OP use during pregnancy was associated with altered cortical activation using functional near infrared spectroscopy (fNIRS) in a small subset (n = 95) of participants at age 16 years.We administered fNIRS to 291 CHAMACOS young adults at the 18-year visit. Using covariate-adjusted regression models, we estimated associations of prenatal and childhood urinary dialkylphosphates (DAPs), non-specific OP metabolites, with cortical activation in the frontal, temporal, and parietal regions of the brain during tasks of executive function and semantic language.There were some suggestive associations for prenatal DAPs with altered activation patterns in both the inferior frontal and inferior parietal lobes of the left hemisphere during a task of cognitive flexibility (β per ten-fold increase in DAPs = 3.37; 95% CI: -0.02, 6.77 and β = 3.43; 95% CI: 0.64, 6.22, respectively) and the inferior and superior frontal pole/dorsolateral prefrontal cortex of the right hemisphere during the letter retrieval working memory task (β = -3.10; 95% CI: -6.43, 0.22 and β = -3.67; 95% CI: -7.94, 0.59, respectively). We did not observe alterations in cortical activation with prenatal DAPs during a semantic language task or with childhood DAPs during any task.We observed associations of prenatal OP concentrations with mild alterations in cortical activation during tasks of executive function. Associations with childhood exposure were null. This is reasonably consistent with studies of prenatal OPs and neuropsychological measures of attention and executive function found in CHAMACOS and other birth cohorts.
View details for DOI 10.1016/j.envres.2023.117756
View details for PubMedID 38016496
Assessment of Surgical Tasks Using Neuroimaging Dataset (ASTaUND).
2023; 10 (1): 699
Functional near-infrared spectroscopy (fNIRS) is a neuroimaging tool for studying brain activity in mobile subjects. Open-access fNIRS datasets are limited to simple and/or motion-restricted tasks. Here, we report a fNIRS dataset acquired on mobile subjects performing Fundamentals of Laparoscopic Surgery (FLS) tasks in a laboratory environment. Demonstrating competency in the FLS tasks is a prerequisite for board certification in general surgery in the United States. The ASTaUND data set was acquired over four different studies. We provide the relevant information about the hardware, FLS task execution protocols, and subject demographics to facilitate the use of this open-access data set. We also provide the concurrent FLS scores, a quantitative metric for surgical skill assessment developed by the FLS committee. This data set is expected to support the growing field of assessing surgical skills via neuroimaging data and provide an example of data processing pipeline for use in realistic, non-restrictive environments.
View details for DOI 10.1038/s41597-023-02603-3
View details for PubMedID 37838752
View details for PubMedCentralID PMC10576768
Short-separation regression incorporated diffuse optical tomography image reconstruction modeling for high-density functional near-infrared spectroscopy.
2023; 10 (2): 025007
Significance: Short-separation (SS) regression and diffuse optical tomography (DOT) image reconstruction, two widely adopted methods in functional near-infrared spectroscopy (fNIRS), were demonstrated to individually facilitate the separation of brain activation and physiological signals, with further improvement using both sequentially. We hypothesized that doing both simultaneously would further improve the performance.Aim: Motivated by the success of these two approaches, we propose a method, SS-DOT, which applies SS and DOT simultaneously.Approach: The method, which employs spatial and temporal basis functions to represent the hemoglobin concentration changes, enables us to incorporate SS regressors into the time series DOT model. To benchmark the performance of the SS-DOT model against conventional sequential models, we use fNIRS resting state data augmented with synthetic brain response as well as data acquired during a ball squeezing task. The conventional sequential models comprise performing SS regression and DOT.Results: The results show that the SS-DOT model improves the image quality by increasing the contrast-to-background ratio by a threefold improvement. The benefits are marginal at small brain activation.Conclusions: The SS-DOT model improves the fNIRS image reconstruction quality.
View details for DOI 10.1117/1.NPh.10.2.025007
View details for PubMedID 37228904
How much do time-domain functional near-infrared spectroscopy (fNIRS) moments improve estimation of brain activity over traditional fNIRS?
2023; 10 (1): 013504
Significance: Advances in electronics have allowed the recent development of compact, high channel count time domain functional near-infrared spectroscopy (TD-fNIRS) systems. Temporal moment analysis has been proposed for increased brain sensitivity due to the depth selectivity of higher order temporal moments. We propose a general linear model (GLM) incorporating TD moment data and auxiliary physiological measurements, such as short separation channels, to improve the recovery of the HRF.Aims: We compare the performance of previously reported multi-distance TD moment techniques to commonly used techniques for continuous wave (CW) fNIRS hemodynamic response function (HRF) recovery, namely block averaging and CW GLM. Additionally, we compare the multi-distance TD moment technique to TD moment GLM.Approach: We augmented resting TD-fNIRS moment data (six subjects) with known synthetic HRFs. We then employed block averaging and GLM techniques with "short-separation regression" designed both for CW and TD to recover the HRFs. We calculated the root mean square error (RMSE) and the correlation of the recovered HRF to the ground truth. We compared the performance of equivalent CW and TD techniques with paired t-tests.Results: We found that, on average, TD moment HRF recovery improves correlations by 98% and 48% for HbO and HbR respectively, over CW GLM. The improvement on the correlation for TD GLM over TD moment is 12% (HbO) and 27% (HbR). RMSE decreases 56% and 52% (HbO and HbR) for TD moment compared to CW GLM. We found no statistically significant improvement in the RMSE for TD GLM compared to TD moment.Conclusions: Properly covariance-scaled TD moment techniques outperform their CW equivalents in both RMSE and correlation in the recovery of the synthetic HRFs. Furthermore, our proposed TD GLM based on moments outperforms regular TD moment analysis, while allowing the incorporation of auxiliary measurements of the confounding physiological signals from the scalp.
View details for DOI 10.1117/1.NPh.10.1.013504
View details for PubMedID 36284602
Deep learning-based motion artifact removal in functional near-infrared spectroscopy
2022; 9 (4): 041406
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 connectivity related to surgical skill dexterity in physical and virtual simulation environments (vol 8, 015008, 2021) NEUROPHOTONICS 2021; 8 (3): 039801
Functional Brain Imaging Reliably Predicts Bimanual Motor Skill Performance in a Standardized Surgical Task
IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING
2021; 68 (7): 2058-2066
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
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
The Effects of Transcranial Electrical Stimulation on Human Motor Functions: A Comprehensive Review of Functional Neuroimaging Studies
FRONTIERS IN NEUROSCIENCE
2020; 14: 744
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
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