The UBC Social Health Lab:
The UBC Department of Psychiatry:
If this sounds like something you are interested in, please send an email to Dr. McLarren at
Very Brief Background and methods:
Language development in children with ASDs tends to be delayed relative to typically developing children, with up to 50% of affected individuals using their first words after 24 months and phrases after 33 months (Spence SJ et al (2006)). In 90% of the typically developing population, the left cerebral hemisphere is critical for speech and
auditory processing (Knecht S et al 2000). Some investigators have found atypical language lateralization in individuals with ASD as assessed by electroencephalography (EEG) (Dawson et al 1986), functional magnetic resonance imaging (fMRI) (Takeuchi et al 2004) and magnetoencephalography (MEG) (Flagg et al 2005). It has not been determined if there is a correlation between delayed language development and the degree of atypical structure and/or function in individuals with ASD. In 2009-2010, our group observed more pronounced atypical lateralization in high-functioning ASD individuals with a history of language delay versus non language-delayed individuals using MEG. Thus, we are comparing the language lateralization of high functioning adults with ASDs with or without language delay, relative to their IQ-matched neurotypical peers. Our hypothesis is that language delay in individuals with ASD is associated with atypical language lateralization and increased involvement of the right hemisphere in semantic processing. Furthermore, this atypical arrangement suggests alternative mechanisms of cognition. This study will clarify and extend our initial findings by looking for convergent evidence using different brain imaging methodologies.
As an extension to our previous MEG work, since 2014 we have been collecting fMRI scans and EEG measurements on subjects as they read sentences. The reading task is designed to elicit differences at the level of semantic comprehension. We have structural MRI scans and diffusion tensor imaging (DTI) tractography for each subject as well, to look for evidence of altered structural connectivity in each group, and to find shared patterns of atypical language lateralization and processing.
Initial processing of these datasets is done using open-source software suites running in Matlab or from the command line in Linux/Unix systems, specifically FSL, Freesurfer, SPM12, EEGlab, and Brainstorm. Recently, we have been moving towards Python-based frameworks to link these processing tools, such as the Nipype suite, and we are open to using some machine learning tools to explore trends in the functional data, for example in Nilearn. We also have access to some bespoke tools written and compiled in C to explore frequency patterns associated with semantic language processing in MEG and EEG data.
Themes of potential 402 projects:
I’ve based these on methodology rather than research question, as I feel we could develop the questions with the students within these frameworks. All of these are open to discussion and development with potential students, depending on interest and ability. Methods and expertise to do these are documented in key references in the literature, in online tutorials, in our lab and neighbouring labs. We can accept two students.
fMRI: the student will learn to read event-related fMRI datasets and develop a method to show effective connectivity networks as matrices. The student may learn to detect different functional networks using principal and independent components analyses. The student may learn to display these networks as heat maps projected onto an anatomical brain. Initial analysis will be on a per-subject basis but can extend to group analyses.
EEG: the student will learn methods of preprocessing EEG data to yield event-related potential (ERP) analysis, principal- and independent- components analyses. The student will learn to distinguish and display induced versus evoked gamma wave responses to sentences. The student may choose to investigate ways to use the data for source reconstruction and beamforming analyses.
MEG: the student will learn methods of preprocessing and automation of preprocessing for event-related MEG datasets using the Brainstorm software suite. The student will use spectral analyses and a tool developed in-house to measure induced versus evoked gamma waves in homonym comprehension. A very able student may choose to investigate metastable brain states associated with comprehension.
DTI and structural: the student will learn to use these scans to reconstruct structural connectivity networks using tools within the FSL and Freesurfer software suites. The student can use data from the fMRI studies to display the structural correlates of functional networks. The student can develop automated processing pipelines in Python/Nipype.
On the highest level, the work brings up interesting questions about how functional networks develop in the brain, how we comprehend language and about mechanisms of cognitive control within the brain. It also addresses a couple of leading theories about the brain basis of autism.