Syllabus database for doctoral courses
-
Startpage
Syllabus database for doctoral courses
SYLLABI FOR DOCTORAL COURSES
Swedish title | Linjär regressionsanalys i neurovetenskap: modellval, implementering, slumpfel och tolkning |
---|---|
English title | Linear Regression Analysis in Neuroscience: Model Choice, Implementation, Analysis Errors and Interpretation |
Course number | 5569 |
Credits | 3.0 |
Notes |
The course meets the requirements for a general science course. |
Responsible KI department | Institutionen för neurobiologi, vårdvetenskap och samhälle |
Specific entry requirements | Basic Course in Medical Statistics (for example KI courses in basic statistics for doctoral students) and being familiar with basic R programming. |
Grading | Passed /Not passed |
Established by | The Committee for Doctoral Education |
Established | 2023-09-25 |
Purpose of the course | The main purpose is for students to understand basic regression analysis principles through a variety of examples with Neuroscience data and help students adapt them to the needs of their research questions using the R language. The course content has been developed to solve regression problems often faced in Neuroscience and all examples will be based on Neuroscience studies. |
Intended learning outcomes | After the course the student is supposed to:
- Be able to explain how linear regression analysis works and decide when regression should/should not be used depending on the research question. - Be able to implement and assess a linear regression model in a basic statistical programming language from scratch but also with packages (This course is not focused on learning R, basic R programming skills are required). |
Contents of the course | The course covers both theoretical and practical regression aspects, with a focus on the application and errors of linear regression modelling.
Special focus throughout the course will be given to the: - Regression optimization, warnings and errors produced and their relation to study design, available data (singularity, autocorrelation, small sample sizes, outliers), and model output (heterogeneous residuals etc). - Search for alternative models that can tackle a scientific question more accurately. - Understanding fixed and random effects. The course focuses mainly on the application of linear regression. Thus, the student will learn how/why linear regression works by using simulated and real datasets from the field of Neuroscience (including brain imaging, neuropsychological assessment, fluid markers). The course is divided in two parts: Part one: 1) Statistical hypothesis testing versus statistical modelling (Introduction). 2) Linear regression. 3) Alternative linear regression models (PCR, PLS, lasso/ridge/elastic net, variable selection models). Part two: 4) Polynomial regression. 5) Introduction to random and fixed effects. |
Teaching and learning activities | Teaching is predominantly in the form of lectures and computer labs. Student activities will include group discussions and presentations, quiz, Q&A sessions, computer labs, and two assignments. The students will receive the time schedule before the course to adjust their working activity accordingly. |
Compulsory elements | Lectures, student group discussion and labs. The course responsible assesses whether and if so, how absence can be compensated. |
Examination | Examination of the course's intended learning outcomes is carried out through a written classroom examination. |
Literature and other teaching material | Recommended course literature:
Faraway, Julian J. Linear models with R. Chapman and Hall/CRC, 2015, ISBN: 978-1-4398-8734-9. |
Course responsible |
Konstantinos Poulakis Institutionen för neurobiologi, vårdvetenskap och samhälle +46704136378 konstantinos.poulakis@ki.se Yrkesvägen 34 14143 Huddinge |
Contact person |
Billy Langlet Institutionen för neurobiologi, vårdvetenskap och samhälle +46762033996 billy.langlet@ki.se Konstantinos Poulakis Institutionen för neurobiologi, vårdvetenskap och samhälle +46704136378 konstantinos.poulakis@ki.se Yrkesvägen 34 14143 Huddinge |