Make MDM.m fully implement MDM about psychtaskframework

[shippy]

Actual properties, per ReadMe.txt:

• 4 repeats for each trial except for the lowest reference level (it's a catch trial), adding up to 78
• An additional trial with reference level in each block -> 80 trials for each domain
• An additional trial with reference level at the start of each block -> 84 trials for each domain
• 8 blocks, 21 trials per block
• Jittered 5s ITI average
• Money symbol in the monetary task.

@rj299, can the first two properties be re-phrased as 4 repeats for each trial that has non-lowest payoff and 2 repeats for each trial that has the lowest payoff? What am I missing?

[shippy]

For future reference, this is the trial generation logic. It presents an interesting challenge for the generateBlocks/generateTrials mechanism (#53).

Front-runner solution is to allow generateBlocks to take as argument an ad-hoc trial table to concatenate with the trials generate through standard means. (Alternatively, we might separate responsibilities completely and pre-generate the trials that we need, then feed them to generateBlocks with hollowed out settings.)