In this unit we write continue working on the linear regression example we have been building up to in the units 2 + 3 of the imaging part.

First, we need to modify the function makeDesignMatrix() that you worked on last time. We want to add 2 more columns to X inside your function:

1. a ramp from -1 to +1
2. a column of ones

These additional columns in the design matrix will capture any slow drift (1st additional column) and mean offset from 0 (2nd additional column). The logic of this will be discussed in class.

Next we will work on a function returnStats() that will allow us to do some statistical significance calculations / parametric stats. The function should be called as returnStats(y, X, c), where

• input arguments are:

  • y is the timeseries data
  • X is the design matrix
  • c is a contrast vector

• finds: beta = X\y; 
using backslash

• or beta = inv(X'*X)*(X')*y;

• finds: modelfit = X*beta;

• finds: r2 - the coefficient of determination

• calculates: t for contrast c