Tuning Curves

Transcript

0:00 – 0:30  [Basic Definition]  A tuning curve is a graph that relates neural activity to a continuous range of stimulus properties.  A typical tuning curve is a bell-shaped plot, and the peak of this plot corresponds to the stimulus that a neuron is maximally responsive to.  For example, some neurons in the primary visual cortex show orientation selectivity and fire most when a line is presented at the correct angle and in the correct part of the visual field.  A tuning curve for this type of cell would plot neural activity on the y-axis and line angle on the x-axis to reveal the angle that generates maximal response.

0:30-2:30  

So in general cells in V1 can be on or off selective. They respond to light turning on or off, same as the retina. They do have center-surrround organization, just like in the retina. But the receptive fields are not round – they’re bar shaped.

In fact, um, it’s not just bars that they respond to, but bars with a certain orientation. Vertical, horizontal at a certain angle. So if you were recording extracellularly from a neuron in V1 and you presented bars of different orientations to the cat, you’d find one particular orientation that made the cell fire the most. Similarly oriented orientations would make the cell fire as well, but probably not as much. And then far away from that orientation, like 90 degrees rotated, you get like nothing.

And now if you plot that relation between the stimulus orientation and the spike rate, and fit a curve to the plotted points, you get sort of a curve with a bump. The bump in the center is the orientation where the cell is firing maximally – the spike rate is the highest. We call this plot a tuning curve because it shows us what kind of stimulus the neuron is tuned to, the one it responds to the most strongly. This is just like a radio that can be tuned to a specific frequency, so that it responds most strongly to that frequency.

We’ll see other examples of tuning curves in future lectures.

We say that cells like this prefer a certain orientation of bar or edge. So we say that they’re selective for a particular orientation or that they have orientation selectivity. Neurons in V1 are orientation selective.

It turns out there is some higher order structure in V1 as well. If you’re doing a recording with an electrode that’s going sort of perpendicular to the cortical layers and you record from a cell here and then another cell here and then another cell here and then another cell here, and so forth. If you do the same experiment where you show bars at all different orientations and then you plot the tuning curves for each cell, all the tuning curves will look really similar. The orientation tuning won’t change as your electrode moves through the tissue.

2:30-3:00  [Parallel Vocabulary] In introductory classes you learned about tuning curves.  But Neuroscience is an interdisciplinary field, so there are many words for this term – For example, when specifically referring to the auditory system, the sound frequency that corresponds to the peak of a tuning curve is called the characteristic frequency and this is the sound frequency that a specific cell is most sensitive to. Broadly speaking, a tuning curve falls within the umbrella category of stimulus response curves, which includes plots that correlate neural activity with any feature for example muscle force or stimulus intensity.

3:00-4:00  [Here’s a real world example]  Detailed knowledge of tuning curves has played an important role in the development of hearing loss treatments. For example, did you know that when hearing loss is caused by damage to the hair cells, it can be treated by implanting microelectrodes in the ear to directly excite the neurons that would have received signals from the hair cells? The device is called a cochlear implant, and it works because within the ear, the neurons that are tuned to specific frequencies of sound are found in predictable locations.  Recall that hair cells within the cochlear synapse with spiral ganglion afferents that in turn project to the brainstem.  Each hair cell is tuned to a specific frequency of sound, and the spiral ganglion afferent that it communicates with is similarly tuned.  This relationship can be demonstrated with a tuning curve that plots sound frequency against neural activity.  By directly exciting the spiral ganglion afferents, a cochlear implant can bypass the hair cells, and generate signals in the spiral ganglion afferents even though the hair cells can no longer function.  Importantly, cochlear implants are more effective for children than for older adults, and this may relate to changes in cognition and the potential for neuroplasticity.

4:00-6:00 [Follow along with this example]

6:00-6:30 [Here are a few readings to help you review]
1) Neuroscience Exploring the Brain (Bear)

• Chapter 11: “The Auditory and Vestibular Systems”

2) Neuroscience Exploring the Brain (Bear)

• Chapter 10: “The Central Visual System”