Course:

Neuroforecasting
Psych 24N-01 (3 units, letter grade)
M 2:30-5:30 PM
01/06/20-03/13/20
Online (canvas.stanford.edu)
Prerequisites: First-year seminar (instructor consent).

Instructor:

Brian Knutson, PhD
Psychology & Neuroscience
Bldg 420, Room 470
Email: knutson'at'stanford.edu
https://stanford.edu/~knutson
Office Hours: W 2:00-3:30 PM (book through rschumm'at'stanford.edu)

Description:

This seminar explores whether brain activity can be used not only to predict the choices of individuals, but also to forecast the behavior of separate groups of individuals in the future (e.g., in markets). Questions include how neuroforecasting might be possible, whether it can add value to other forecasting tools (e.g., traditional measures like behavioral choice and subjective ratings), and when it might extend to different aggregate scenarios. The course is ideal for students with interdisciplinary interests who would like to extend neural predictions about individual choice to group choice, and who plan to apply this knowledge in their future research.

Text:

(readings are linked to the syllabus below)

Schedule:

  • Classes 1-2 provide minimal background in neuroscience and psychology.

  • Classes 3-7 include discussion of relevant research and findings.

  • Class 8 focuses on students' research presentations of proposed projects.
    • Date Theme Readings
    21.06.21 Defining Neuro | Defining Forecasting Samanez-Larkin & Knutson (2015) (pp. 279-280) | Walonick (1993) / Clements et al (2002)
    21.06.28 Data Acquisition | Data Analysis Kable (2012) | Dawes (1979) / Yarkoni & Westfall (2017)
    21.07.05 (Independence Day) --
    21.07.12 Superforecasting | Prediction with physiology Mellers et al (2015) | Schoemaker & Tetlock (2016) | Wang & Minor (2008)
    21.07.19 Prediction with FMRI Ariely & Berns (2010) | Knutson et al (2007) / Tusche et al (2010)
    21.07.26 Forecasting with EEG Hakim & Levy (2018) | Telpaz & Levy (2015) / Barnett & Cerf (2017)
    21.08.02 (BK out) --
    21.08.09 Forecasting with FMRI Berkman & Falk (2013) | Falk et al (2012) / Scholz et al (2017) / Genevsky et al (2017)
    21.08.16 Potential & limits | Ethics Knutson & Genevsky (2018) | Venkatraman et al (2015) / Tong et al (2020) | Varan et al (2015)
    21.08.23 Summary | Presentations

    Requirements:

    Grading:

    Notes (30) / Critique (10) 40%
    Outline (10) / Presentation (5) / Proposal (25) 40%
    Attendance (10) / Attention (10) 20%


  • Notes: In Classes 2-7, students should prepare a half-page summary and one question raised by each of the assigned papers ("Notes"). Summaries and questions are due electronically midnight the Sunday before each class, with no extensions allowed (but see the grading policy below). The goals of these assignments are to stimulate critical thinking about relevant research and to ensure active participation in class discussions.

    Summaries should address the following points:
    1. For reviews, what were at least three main points of the review, are there any you questioned, and why?
    2. For experiments, what was the independent variable, dependent variable, prediction, and conclusion?
    3. Did the findings justify the conclusion? Why or why not?
    4. What is one way you might improve or extend the work?

  • Critique: Before the third class, you and your partner should submit a one-page critique of a company that is trying to use physiological data to forecast aggregate choice. The goal of this exercise is to develop the ability to empirically evaluate commercial claims of forecasting. Submit your critique electronically by midnight on 21.07.09 (July 09, 2021). The critique should address:

    1. What did the company claim they could forecast, and with which physiological data?
    2. Was there any peer-reviewed published evidence that the company could use their proposed method to forecast what they claimed? Did this evidence actually support the claim? (include citation)
    3. Was there any evidence of added value (i.e., that the proposed method could do better than or add value to other methods)?
    4. Was the claim made based observational, explanatory, or predictive evidence?

  • Outline / Presentation / Proposal: This assignment involves developing a concise proposal to conduct an experiment relevant to neuroforecasting. The goal of this assignment is to learn to navigate the process of innovating and designing fundable interdisciplinary research.
    Students should first electronically submit an outline for the proposal by midnight on 21.08.13 (August 13, 2021) for early feedback (one page, bullet point ok). The outline should include independent variable, dependent variable, prediction, potential confounds, and references to 3-5 relevant sources.
    During the final class, each student will briefly present their proposal outline in powerpoint format (5 min) and receive additional feedback (2 min; order to be randomly determined). Proposal outline slides are due via email at midnight on 21.08.20 (August 20, 2021).
    Final proposals should be in a brief ~5 page format (example to be linked here). The final proposal is due at midnight on 21.08.25 (August 25, 2021)


  • Attendance / Attention: Valid reasons for absence include: (1) arrangements made in advance with the instructor to participate in a Stanford sanctioned activity (e.g., athletic competition), (2) a medical condition that requires the treatment of a physician, or (3) the death of a close family member.
    "Attention" is essentially psychological participation. Don't use devices in class, as research indicates that they impair learning and decrease grades (unless explicitly required for exercises).


  • Guidelines: Plagiarism is considered academic theft -- it cheats the creator, the instructor, and the student. I can and will identify it, which can result in a failing grade.


    • https://stanford.edu/~knutson/nfc/nfcsyll.html (updated: 21.06.21)