The August 2015 issue of Minnesota Medicine includes an article by Carmen Peota about BSC researcher Dr. Lisa James and her three-year study of post-traumatic stress disorder (PTSD) and female veterans. "A Focus On Female Vets" describes Dr. James' interest in "defining the neural signature of PTSD for women and finding a genetic basis for why some women might be more resilient than others."
Margaret (Maggie) Mahan, Chelley Chorn and Apostolos Georgopoulos have published "White Noise Test: detecting autocorrelation and nonstationarities in long time series after ARIMA modeling" in the Proceedings of the 14th Python in Science Conference (SciPy 2015).
"In this paper, we presented an expansion on the Box-Jenkins methodology to ARIMA modeling ... Overall, using our approach, an investigator can perform ARIMA modeling and evaluate candidate models with ease for large datasets and datasets containing long time series." (Mahan et al., 2015)
Time series analysis has been a dominant technique for assessing relations within datasets collected over time and is becoming increasingly prevalent in the scientific community; for example, assessing brain networks by calculating pairwise correlations of time series generated from different areas of the brain. The assessment of these relations relies, in turn, on the proper calculation of interactions between time series, which is achieved by rendering each individual series stationary and nonautocorrelated (i.e., white noise, or to "prewhiten" the series). This ensures that the relations computed subsequently are due to the interactions between the series and do not reflect internal dependencies of the series themselves.
An established method for prewhitening time series is to apply an Autoregressive (AR, p) Integrative (I, d) Moving Average (MA, q) model (ARIMA) and retain the residuals. To diagnostically check whether the model orders (p, d, q) are sufficient, both visualization and statistical tests (e.g., Ljung-Box test) of the residuals are performed. However, these tests are not robust for high-order models in long time series. Additionally, as dataset size increases (i.e., number of time series to model) it is not feasible to visually inspect each series independently. As a result, there is a need for robust alternatives to diagnostic evaluations of ARIMA modeling.
Here, we demonstrate how to perform ARIMA modeling of long time series using Statsmodels, a library for statistical analysis in Python. Then, we present a comprehensive procedure (White Noise Test) to detect autocorrelation and nonstationarities in prewhitened time series, thereby establishing that the series does not differ significantly from white noise. This test was validated using time series collected from magnetoencephalography recordings. Overall, our White Noise Test provides a robust alternative to diagnostic checks of ARIMA modeling for long time series.
BSC researcher Dr. Sofia Sakellaridi has a paper appearing in the August 2015 issue of Experimental Brain Research entitled, "Neural mechanisms underlying the exploration of small city maps using magnetoencephalography". Along with her colleagues Peka Christova, Vassilios Christopoulos, Arthur C. Leuthold, John Peponis and Apostolos P. Georgopoulos, Dr. Sakellaridi "conducted a novel brain imaging experiment to test the hypothesis that a network of cortical regions is involved in the processing of spatial information acquired during exploration to make a decision. We recruited 10 subjects and asked them to explore small city maps exemplifying five different street network types (i.e., regular, colliding, curvilinear, cul-de-sac, and supergrid) to build a hypothetical City Hall, while neuronal activity was recorded continuously by 248 MEG sensors at high temporal resolution."
The neural mechanisms underlying spatial cognition in the context of exploring realistic city maps are unknown. We conducted a novel brain imaging study to address the question of whether and how features of special importance for map exploration are encoded in the brain to make a spatial decision. Subjects explored by eyes small city maps exemplifying five different street network types in order to locate a hypothetical City Hall, while neural activity was recorded continuously by 248 magnetoencephalography (MEG) sensors at high temporal resolution. Monitoring subjects' eye positions, we locally characterized the maps by computing three spatial parameters of the areas that were explored. We computed the number of street intersections, the total street length, and the regularity index in the circular areas of 6 degrees of visual angle radius centered on instantaneous eye positions.
We tested the hypothesis that neural activity during exploration is associated with the spatial parameters and modulated by street network type. All time series were rendered stationary and nonautocorrelated by applying an autoregressive integrated moving average model and taking the residuals. We then assessed the associations between the prewhitened time-varying MEG time series from 248 sensors and the prewhitened spatial parameters time series, for each street network type, using multiple linear regression analyses.
In accord with our hypothesis, ongoing neural activity was strongly associated with the spatial parameters through localized and distributed networks, and neural processing of these parameters depended on the type of street network. Overall, processing of the spatial parameters seems to predominantly involve right frontal and prefrontal areas, but not for all street network layouts. These results are in line with findings from a series of previous studies showing that frontal and prefrontal areas are involved in the processing of spatial information and decision making. Modulation of neural processing of the spatial parameters by street network type suggests that some street network layouts may contain other types of spatial information that subjects use to explore maps and make spatial decisions.
BSC investigators Vassilios Christopoulos, Angeliki Georgopoulos and Apostolos P. Georgopoulos have published their paper entitled,"The effect of apolipoprotein E4 on synchronous neural interactions in brain cultures" in the April 2015 issue of Experimental Brain Research. To their knowledge, this is the first study of the effects of apoE4 on neural network function in vitro. ApoE4 has been associated with various aspects of brain function and disease. The mechanisms of action of apoE4 in the brain are only partially understood and encompass various levels of reference. At the gross disease level, apoE4 is a known risk for Alzheimer’s disease, is involved in early onset of Alzheimer’s disease neuropathology in Down’s syndrome, adversely affects the sequelae of traumatic brain injury, affects susceptibility, clinical type and progression rate in multiple sclerosis, and is associated with higher symptom severity in posttraumatic stress disorder (PTSD).
Apostolos P Georgopoulos
Our newest colleague is native Minnesotan Rachel Johnson, who will be working with Dr. Lisa James as a Program Support assistant. Rachel holds a BA in Linguistics, Russian and German and an MA in ESL from the University of Minnesota. After graduation, she taught English to immigrants, refugees and native-born adults with special learning needs. More recently, she has been managing a school for adult English Language Learners in St Paul's North End. Rachel will be pursuing a PhD in Cognitive Science at the UofM under the supervision of Dr. Apostolos Georgopoulos.
BSC researcher and University of Minnesota Cognitive Science PhD student Nicole Scott successfully defended her doctoral thesis "Cognitive and Neural Correlates of Processing Spatial Relations by Humans" on the morning of August 31st, 2015. In attendance were thesis committee members Maria Sera (adviser), Apostolos Georgopoulos (co-adviser), Matt Chafee and Jeanette Gundel, along with many friends and colleagues
Cognitive and Neural Correlates of Processing Spatial Relations by Humans
Human cognition has long been thought to exceed that of other animals; however, what it is that makes humans "so smart" continues to be questioned. Gentner argues that language and relational reasoning together elevate human cognition and she takes a developmental approach to support her theory. This project takes a similar approach to Gentner's: I examined the relationship between language and relational reasoning in children, specifically as they are learning the relational terms for right and left as compared to relations for terms that they already know (i.e., above/below).
What sets this project apart from Gentner's work is that I also looked at the effect of lateralization on children's performances as well the neural mechanisms underlying these same relational judgments in adults. We know some of the neural mechanisms underlying relational reasoning in humans and monkeys. We do not, however, know whether one set of relations (i.e., above/below) should be advantaged over the other (i.e., right/left) in both verbal and nonverbal modalities.
To answer this question, I have developed a set of tasks to examine the following queries:
1) how verbal and nonverbal knowledge of above/below/right/left develops from 5 years to 10 years of age,
2) whether verbal knowledge aids performance on a nonverbal task that requires judgments of these relations,
3) whether strength of handedness promotes either verbal or nonverbal performance, and
4) what the neural correlates of these judgments are.
"Neural communication in posttraumatic growth" is the most recent publication by BSC researchers Samantha L. Anders, Carly K. Peterson, Lisa M. James, Brian E. Engdahl, Arthur Leuthold and Apostolos P. Georgopoulos. In this magnetoencephalography (MEG) study, they looked at posttraumatic growth (PTG), or positive psychological changes following exposure to traumatic events. The article is featured in the April 17 2015 issue of Experimental Brain Research.
Posttraumatic growth (PTG), or positive psychological changes following exposure to traumatic events, is commonly reported among trauma survivors. In the present study, we examined neural correlates of PTG in 106 veterans with PTSD and 193 veteran controls using task-free magnetoencephalography (MEG), diagnostic interviews and measures of PTG, and traumatic event exposure. Global synchronous neural interactions (SNIs) were significantly modulated downward with increasing PTG scores in controls (p = .005), but not in veterans with PTSD (p = .601). This effect was primarily characterized by negative slopes in local neural networks, was strongest in the medial prefrontal cortex, and was much stronger and more extensive in the control than the PTSD group. The present study complements previous research highlighting the role of neural adaptation in healthy functioning.
Extending results from their MEG studies of 2010, Brain Sciences Center researchers Peka Christova, Lisa M. James, Brian E. Engdahl, Scott M. Lewis and Apostolos P. Georgopoulos have published the paper "Diagnosis of posttraumatic stress disorder (PTSD) based on correlations of prewhitened fMRI data: outcomes and areas involved" in the journal Experimental Brain Research. Christova and her team successfully classified PTSD and control subjects using neural correlations from prewhitened resting-state fMRI data with 93.3% accuracy.
Successful diagnosis of PTSD has been achieved using neural correlations from prewhitened mag- netoencephalographic (MEG) time series (Georgopoulos et al. in J Neural Eng 7:16011, 2010. doi:10.1088/1741- 2560/7/1/016011; James et al. 2015). Here, we show that highly successful classification of PTSD and control sub- jects can be obtained using neural correlations from pre-whitened resting-state fMRI data. All but one PTSD (14/15; sensitivity = 93.3 %) and all but one control (20/21; speci- ficity = 95.2 %) subjects were correctly classified using 15 out of 2701 possible correlations between 74 brain areas. In contrast, correlations of the same but non-prewhitened data yielded chance-level classifications. We conclude that, if properly processed, fMRI has the prospect of aiding significantly in PTSD diagnosis. Twenty-five brain areas were most prominently involved in correct subject classification, including areas from all cortical lobes and the left pallidum.