UCSB BRAIN

Curriculum

Overview of differences computational science

rweissle@uoregon.edu

Rachel Elizabeth Weissler

I’ll focus on overview of sociolinguistics and how it relates to computational neuroscience:

Linguistics is the scientific study of language and how people use language. Linguistics is not the study of how to “speak good” or “talk well,” as ‘good’ and ‘well’ are not scientifically valid concepts. Linguists look at how language is spoken by humans, track variation and changes throughout history with relation to language, and do analyses based on the data – the language output. The field of linguistics is broad. The field of sociolinguistics is the study of language in relation to social factors, including differences of regional, class, and occupational dialect, and gender differences.

So how does this relate to neurocomputation?

Well, there is bountiful evidence from neurolinguistics (people who study how the brain processes language) showing that people invoke prediction during sentence processing (Van Berkum et al. 2008, Hanulikova et al. 2012, Kutas et al. 2014). This means when we are listening to someone talk, our brain is one step ahead of us, trying to predict what will come out of that person’s mouth next, based on the context of the conversation. Electroencephalography (EEG) is one method used by neurolinguists to look at this predictive processing. EEG measures electric potentials that are generated by tens of thousands of cortical neurons using electrodes placed on the scalp. Averaging the EEG signal that is recorded to multiple instances of a specific perceptual speech event reveals systematic voltage changes associated with the cognitive processes elicited by that event, called the Event-Related Potential (ERP). This neurocomputational tool is useful to linguists because it allows researchers to passively monitor neural activity which reflects implicit and on-line linguistic judgements, including the so- cial expectations of listeners. It also illuminates for researchers when, in real time, expectations on the part of the listener may be violated during processing, indicating that prediction is taking place.

A great deal of prior work has revealed ERP signatures for semantic and syntactic violations (Luck 2005). For example, if something is semantically unexpected in a sentence string, a listener will exhibit an increase of negative voltages over the central scalp that peaks around 400 milliseconds after word onset (the “N400”). Let me give you an example. The sentence “I like my coffee with cream and sugar” is a pretty normal sentence. You know what it means, I talked about coffee, and as the sentence unfolded, you probably could have predicted that i was going to say sugar from what you know about coffee drinking. However, if I were to say, “I like my coffee with cream and shutters,” that might shock you. Both sugar and shutters have the same initial consonant sound, and they’re both nouns, but shutters doesn’t belong in the coffee context. When we hear shutters instead of coffee, our brain shows a semantic neurological response of surprisal, because our predictions during processing were violated. Thus, EEG is a great tool for people interested in the social information that listeners hold when they’re listening to others speak, as we can learn more about what the brain is capable of recognizing and processing.

Potential Student Research Plan for students working with Weissler * *All Student tasks should be completed prior to that week’s meeting:

Week 1

Rachel	Students
– Meet with research assistants, set expectations and goals	– Attend intro meeting

Week 2

Rachel

Students

– discuss linguistics research

– discuss neurolinguistics

– find a time that works to meet weekly for 30 min

– download/buy *experiment builder)

– download Praat

– Read Luck EEG Technique here

– write up any questions you

Week 3

Rachel

Students

– discuss questions

-Training in audio splicing

– read Weissler & Brennan 2019

– Read about EEG Experiments here

– read Van Berkum et al and Hanulikova et al.

– write up paragraph for questions and comments about each paper and EEG Experiments

Week 4

Rachel

Students

– discuss alternatives due to lab not opening

– gain access to MatLab

– Watch fieldtrip toolbox videos to learn about EEG analysis

– Note any roadblocks and bring them to meeting

Week 5

Rachel	Students
-send audio files and data files	– Work through files in MatLab, note any issues

Week 6

Rachel	Students
– Discuss the data in matlab and how to do preprocessing	– Look through Multi-Image Analysis GUI (MANGO) and watch video tutorials

Week 7

Rachel	Students
– brainstorm project types	– Practice preprocessing data in MatLab

Week 8

Rachel	Students
-discuss trigger timing measurements and its importance in EEG research	– practice creating trigger times with Praat and Excel with google sheets

Week 9

Rachel

Students

– other studies to be a part of

– paid opportunities

– closing remarks

– come to meeting prepared to discuss what you’ve learned over the course of the active research time, and what skills you believe you’ll take with you