A branch of psychology whose main purpose is to determine what the neural bases for cognition are.

This entails four primary goals:

1. Determining how the brain performs neural computations.

2. Determining how these neural computations are used to implement cognitive processes.

3. Determining where, if anywhere, in the brain these processes are carried out.

4. Determining how these systems interact.

Four vital features of how neurons fire:
 

1. Neurons have excitatory and inhibitory connections to each other.

2. A neuron only fires when its net charge is above the threshold of excitation.

3. A neuron’s action potentials are always the same magnitude.

4. The strength of a neuron’s activation is measured by the firing rate.

In order for a model of a cognitive process to be truly plausible, one must be able to conceive of a mechanism by which the model can be implemented with the constraints imposed by the properties of neural computations.
 
 
 
 

Unfortunately, we don’t understand the details of neural computations yet to accurately constrain these choices. For example, how do two neurons know to grow connections to each other?

Ideas and issues about generic mechanisms:

1. Hebbian learning: If neuron A has an excitatory connection to neuron B, and neuron B fires immediately after neuron A, then increase the strength of the connection between neuron A and neuron B.

Problem is, Hebbian learning cannot explain how learning happens across multiple layers of neurons, such as in vision.

2. Modular vs. Distributed systems: Are there specific pieces of the brain that are responsible for specific processes, independent of the other pieces? Or is the implementation of cognitive processes distributed across the brain in a more interactive fashion?

Five primary methodologies:

1. Post-mortem analysis: Find someone with some interesting problem, such as an inability to produce proper speech, wait for them to die, and then look at what part of their brain is missing.


Examples: Broca’s and Wernicke’s areas; Phineas Gage

2. Human lesion study: With today’s imaging technologies (MRI, PET, etc.), we can see where a person’s lesion is before they die. Thus, we can administer a finer grain of cognitive tests to determine their cognitive deficiency.


Examples: H.M.; optic aphasia
 
 

3. Monkey lesion study: Unlike in people, we can control where we put a lesion in a monkey and then look at what tasks the monkey is no longer able to perform.


Examples: Newsome & Paré (on Thursday)
 
 

4. Single-cell recording: We can also record the activity of individual cells in the monkey’s cortex and see how they respond to various stimuli or task demands.


Examples: Receptive fields in visual cortex

5. Normal brain scanning: We can use functional MRI to look at brain activity in normal subjects to determine what parts of the brain are more active during particular tasks.


Examples: mental imagery; object recognition

6. Neural stimulation: Stimulate part of brain and see what twitches.