LECTURE 8 Flashcards
Explain the concept of an ROC curve, how its shape depends upon the properties of the underlying decision space, and describe one example study that illustrates how this signal detection approach can be useful in neuroscientific experiments.
An ROC (Receiver Operating Characteristic) curve is a graphical representation that evaluates the performance of a binary classifier system across various classification thresholds. It plots the hit rate on the y-axis against the false alarms rate on the x-axis. Depending on different d’ prime values our curve changes shape the lower the d’ value the closer the curve will be to a straight diagonal line (d’=0).
A paper studying monkeys that had to make decision about line orientation. Then they measured data of one neuron in V4 and then they can take the average of all samples across trials. from this data we can plot the histograms for both the orientations, and this will give us the decision space. With this data from one neuron, we would like to know how accurate our neuron would be on the task. So, what they did is they made an ROC curve and the area under this curve gives the upper bound of accuracy.
Describe a study that shows that the accumulated evidence as assumed by diffusion decision models can be found in neuronal recordings. (Make sure to explicitly explain which characteristics of the results make us believe that this is indeed a signal that reflects the accumulated evidence as assumed by the models)
Monkey task: Decide if dots are moving left or right.
Manipulated how many dots are moving in the same direction
Recorded neuron in MT but here the firing rate of the neuron doesn’t accumulate so here there is no evidence for the accumulation.
However, another area in the parietal cortex has this evidence accumulation the firing of the neurons goes up overtime and the more motion strength there is the faster this goes up. Also when you plot the data in another way where 0 means the time where the monkey makes a response we can see that all the signals reach the same strength of before the monkey makes the decision.
