BP Post exam 1 - Notes from after midterm PDF

Preview

Summary

Basic AuditionBasic Principles of soundAn attempt to make inferences about distant events Ex: rustling leaves - Pick up on the infor and make an inference about where and what it is - PROBLEM: the nature of this information is so minimal relative to the actual event, there are endless possibilities ...

Description

Basic Audition Basic Principles of sound An attempt to make inferences about distant events Ex: rustling leaves - Pick up on the infor and make an inference about where and what it is - PROBLEM: the nature of this information is so minimal relative to the actual event, there are endless possibilities so we need a system capable of doing these things Ex: professor says words that you've never heard before and even though they’re unfamiliar, we can still process them and in his voice. How do we generate a response or transduction of auditory information to do this? Basic Principles of Sound Sound is a wave, a series of compressed air and then decompressed air that travel as a result of collisions in the real world Intensity Frequency Timbre Intensity: amount of vibration produced by sound - Perceived as loudness - Intensity refers to amplitude (height) - High amplitude = loud, - Decibel: logarithmic scale of sound intensity - Every 1 unit change is 10 unit change in amplitude - 160 dB= instant perforation Frequency: how often the wave oscillates - Perceived as pitch - Hertz- cycles per second - Low frequency = lower pitch (less likely to be interfered with and travel distances better) - Pure tone: only one frequency (tuning fork) - Noise: combination of waves, do not regularly repeat - We don’t use really high or low frequency sounds Complex Sounds: Have a clarinet playing middle C Fundamental Frequency: lowest frequency determines note/pitch

-

Waveform of clarient: more complex due to overtones. THere are various oscillations introduced by the instrument Instruments have different overtones Harmonics: determine timbre/quality PROBLEM: How do we know it’s a C on different instruments? - We decompose waveforms into its fundamental frequency

Peripheral Functional Neuroanatomy of audition Tympanic membrane, ossicles, bony labyrinth, cochlea Series of moving bones held together by ligaments → overtime, they degrade Outer Ear: pinna, auditory canal - Collect, focus, and channel sound into the canal - Localize sound Middle Ear: - Tympanic membrane - Ossicles (malleus, incus, stapes) - Oval window Inner Ear: - Semicircular canals - cochlea Transduction of sound into neural signals As the wave makes its way through the canal, it meets the tympanic membrane which vibrates in the same way the air is. As it vibrates, it is connected to a series of bones. The Malleus is directly connected which interacts with the Incus which interacts with the stirrup. It then pushes against the oval window. The window then dives access to the bony labyrinth. The oval window presses into liquid in the cochlea, introducing vibration into the endolymph. It then goes through the cochlea and back out. - This is a series of different transduction events - Oscillation is amplified within bone interactions - This movement becomes the electrical activity Evolution: this system is an amplifier designed to emphasize small vibrations in the air so they can be effectively transduced into the cochlea. The movement of the bones is larger than in the air.

Inner Ear Cochlea: fluid filled organ containing receptors that respond to vibrations in the inner ear 1. Vibrations travels up the tympanic canal 2. Down the vestibular canal As the vibration moves, it moves the membrane that separates the 2 canals. It causes the basilar membrane to move up and down. Now these alternative waves in the air are up and down movement in endolymph which become movement in the basilar membrane. On the BM is the organ of Corti (bunch of hair cells that are attached to the BM and pressed against the tectorial membrane. As the membrane flexes, the hairs will be pressed and released). As hairs are pressed up all of the hair cells will bend which opens the ligand gated channels. THey open, depolarize and cause a series of AP. When they are released, the AP reduces. What ends up happening is an alternating series of a lot of AP and not a lot. The sound becomes electrical activity.

The cochlea

Evolution: Sound Frequencies are translated by the BM - Designed to accomplish some math - Decompose sound into separate pure tones that are present - Cochlea is efficient in packing in a lot of BM into a small area. If it were to be unwrapped and laid out flat, the BM is not uniformly the same across its length. What we have now is an oscillation introduced at the point of the oval window so now you have the movement of the liquid as it makes its way to the apex. As it does, it creates movement in the BM, but the BM has a

different level of stiffness across its length. So at it’s base it is very narrow, stiff and thick and that means when a high frequency sound comes in, it is much more likely to cause a vibration in the BM at the base than anywhere else. In opposition, the apex is wide, thin, and flexible making it more responsive to low frequency sounds. What happens over the entire length of the BM is different frequencies are going to cause different sections of the BM to be vibrated. The particular hair cells at a point are indicative of the presence of a 1000 Hz oscillation. Because this complicated sound has a bunch of different frequencies within it, now the complicated sound causes vibrations in distinct parts of the BM and the firing is distinct as well. Now what the BM has done is called a Fourier transform which is where you take a complicated waveform and break it down into its fundamental frequency and all of its overtones within it. This was a mathematical principle discovered by fourier in 1822 but it’s been functioning in the ears of mammals for millions of years. The responses that came out of the ear before were even in ECA. The vibrations of the BM takes a waveform and the rate of firing the hair cells captures the amplitude of those frequencies. In the rate code of hair cells, you have a decomposed sound. This is the first order trick that allows you to process new sounds. Now we have a common space directly captured in the electrical response in the cochlea into which all sound can be placed. And therefore all sound can be compared. The similarity of one to another to another can easily be derived because you now compare the fundamental frequency. How do we know we’re playing a C, well training but at some level, you pull this trick off in your ear before you get to electrical activity.

Central Mechanisms of Audition (what happens once we're into electrical activity?) Pathways 1. Transduction and responses in the inner ear 2. Electrical response in hair cells (bundled). The cells release glutamate which causes the auditory ganglion cells to fire. Those axons make their way out of the ear via the auditory nerve through and into the VCN and DCN in your brainstem (just at the level of medulla). At that point some info is going to cross over. Infor from the right ear makes its way into the right cortex, a little crosses into the right side of the brain and a little to left. Majority of it will stay on the same side and move into inferior colliculus which lies below the superior colliculus. From there it goes to the

medial geniculate nucleus which lies medial to LGN and into the primary auditory cortex (on the inferior bank of the sylvian fissure). As it goes through each one of these connections it's going to be slightly perterred and certain things are going to be calculated to allow for certain kinds of more complicated behavior. The first thing we have at the level of the cochlea is a version of topography. It's laid out in terms of frequency. Adjacent parts of the BM (similar hair cells embedded within it respond to similar frequencies). That is a topographical representation in a frequency domain. If we look at the different stages of this processing sequence we see that same organization recapitulated. Just as in somatosensory domain, we see the layout of the body recap. In the sensory homunculus. Tonotopy - Tonotopy is maintained throughout the pathway to the cortec (topography) Here in the IC, we can see another map of the frequency span. → smooth map As we move out way into PAC, and map out which frequencies produce the strongest responses, we see a smooth map. A topographic representation that is a recapitulation of the sensory surface which in the case is the BM. Just as we saw the recapitulation of the sensory surface for somatosensation which is the layout of the skin across the body.. The representation of the visual field is the layout retina and it is recapitulated again in the SC which is recapitulated again in early visual cortex. In each case, the layout of the response is directly related to the layout of the sensory surface. This is the first order representation of the outside world. You have an evolved sensory surface and then that sensory surface is In the representations within the primary cortices. As we move outside primary cortices, and into other parts of the brain, this relationship breaks down. In these early areas we see the sensory surface again. Sensation vs Perception Sensation: acquisition of sensory information → what occurs at the sensory surface - decibels: physical quality of the sound stimulus - Physical sensation of pain Perception: interpretation of this information → what you experience - Loudness - Emotional reaction to pain Ex: you have a loud sound that can be described in decibels and it’s so loud that it hurts. We can describe the pain in terms of the response of the somatosensory cells which are innervating the ear. But your perception is different. You perceive loudness which is related to decibels but it’s not the same. Even before info leaves the ear, it’s been broken down by the BM transform. In

doing so, as we go through each step, what is actually happening is different from the actual sound. In many cases it’s adaptive and particular elements of the sound will be emphasized/deemphasized. Some things are not going to be fully transduced. This is a good distinction between sensation and perception.

Localizing sounds in space

-

-

-

-

-

At the most basic levels (brainstem response: IC, medulla nuclei) a lot of the processing is not about the content, but about localization. Figure out where the sound is coming from. The original purpose is to detect unexpected noises and move you towards prey and away from danger. To do that you need to know where the sound is and where it's coming from. Problem is there’s nothing in the sound itself that indicates location. You get amplitude and frequency. How do you figure out where things are in space? You use the fact that you have two ears. The wave of sound reaches right ear .5 ms faster than your left. One purpose of the crossover, is to take in account for the lag. Your system knows the width of your head and can determine what angle a sound is coming from. The asynchrony between sound reaching the two ears signals the angle of the sound relative to the head. There are some ambiguities. This will not help you with sounds in front of behind your head or above or under because they reach the ears at the same time. The reason you have the weirdly shaped ear (pinna) is because it introduces a distortion into sound depending on the direction its coming from. The asynchrony is a really good cue and the pinna introduces small distortions. In doing so it gives a clue to direction because distortions are different no matter what direction the sound is coming from because the pinna is not symmetric. Before the sound enters the ear, a little info is added from the pinna as to direction.

Localization vs Language

Dichotic listening is related to hemisphere lateralization for language.

Ba vs fa - In both clips you hear va, but what's happening is his mouth looks like it’s changing the sound. - The only difference between the sound is the shape of the mouth and so there is enough info to process but vision creates an auditory experience - You construct your reality with guesses (visual information is more direct, richer) the difference is much larger in visual - This all happens before perception. When you think about consciousness and conscious perception, the one point that is true is that you have conscious access to the constructed perception you have of the world. A bunch of heuristics blended together in adaptive clever ways through experience and evolution, to color the things we see. There is an element of construction in vision and audition (guessing). A lot is devoted to inferring and processing events though these two senses. There is a fundamental constructive experience where a lot of what you see is a product of low level guesses placed there by evolution and experience but some of it relates to implicit bias and eyewitness testimony.

Attention: Behavioral Measures Visual Search -

Attention is like a filter The visual world is too complicated and there is no way to simultaneously process all

bits of information that comes in, so we gave to reduce complexity, The task of finding waldo is very complex. Because there's a lot of stimuli that match some part to what you’re looking for. - We have to do something to decrease the complexity by looking at one portion of the image and shift over until you cover the whole image. That leads to classic models that are modular accounts of visual attention. Classic Models are Modular Accounts Theoretical account goes something like this: you have visual information entering the brain and before attention engages at all, you get preattentive processing, producing a simple description of the visual world. In the case of visual its like a very basic, non detailed description of the for example spatial locations in the eye so what is the space on the retina I’m interested in searching is one way to think about it. You can only do this really basic processing before engaging attention. To do more complex, need to engage attention, - Attention is a filter, which selects a subset of the available information for access to “further processing”. It takes this vast amount of information and selects a subset, that subset is the info that makes it up into awareness and gains access to the resources needed to do that fine grain processing. This account is similar to searching for waldo. - These accounts are also called modular accounts. They come from older traditions in which you propose there are stages in which one stage only interacts with one other stage in terms of what is produced. - This Preattentive Processing does this simple basic map of location and salience like brightness, produces the map and the output is the map that is embedded into the attention system. Which looks across the map and selects subsections to allow awareness.. These accounts were largely formed before we had a good understanding of how the neural substrates actually functino. - Attention is localized and modular, with a single filter serving as the gateway between sensation and perception (or conscious awareness) - They make predictions about what happens when we get to cortical debate. The purpose of all of these accounts are to describe complex behavior. The primary source of complex behavior is the cortex. - one of the most direct predictions is what is going to happen with damage. If we lesion a part of the brain and when it occurs, it interferes with the access and detail that awareness has to incoming stimulus information. -

Lesions should specifically impact attention - There should exist patients with damage to an atten area that will interfere with awareness but not basic processing - Turns out that this prediction hold to a degree. When a certain area of the brain is damaged, we get neglect. Neglect - Behavioral disorder where individuals will ignore half of space - This kind of brain damage can be seen in memory too. Took people with neglect to stand in the paza and patients had to recall buildings facing a direction. They only recalled the buildings on one side and when turned around recalled the other side

-

Posterior parietal cortex damage interferes with the ability of conscious awareness to access certain parts of the visual scene. - Thus can find lesions that interfere with awareness. Some processing is intact - But some processing is still occurring. - Presented houses, one on top of the other, one was on fire. The right side of the house was the same. Asked patients if there was anything different. They say the houses are the same, but when asked which they’d rather live in, they often choose the top house even though there is no awareness of burning flames. Its clear that the basic processing of houses is intact and they can clearly choose one house or the other, but that doesn’t come through in the low level question. They say they’re the same but show a high level preference for one house over the other. This suggests that this processing is capable of aiding in high level opinion. Result raises a key question: - How basic is preattentive processing if it includes the formation of preferences? - How extensive is preattentive processing - This resulted in the field asking how do we investigate preattentive processing mechanisms in detail? - The only way is using more advanced behavioral stuff - Now going to look at how we used detailed behavior to understand these mechanisms

Behavioral methods Basis of modern psychology - We test our hypotheses against empirical data to produce genera\izable knowledge and new theories Quantifying behavior - Behavior reduces to accuracy: how often does a participant make the correct judgement in particular contexts, or proportion of times a participant chooses a certain even - And reaction time: how long does it take for them to correctly respond, measure of efficiency of mechanism and how quickly it can find the right answer - Only used with correct response if there is a correct/incorrect - Example: testing accuracy - Get a kiddie pool and fill with opaque liquid in order to obscure stuff inside. Hidden inside is a submerged platform. A rodent standing on it will keep its head above water. Drop a rodent and observe what route they use to find the platform. 1st it swims to find something and explore where it eventually finds the platform. Every additional trial results in increase in efficiency. Can

-

take the length of the path as a measure of learning. That is your accuracy measure. How does it do this? The answer is it is using cues from the route. Where the platform is relative to other objects in the environment. If you were to rotate the pool, the rat becomes poor at finding it. Now can figure out what parts of the brain are responsible for this process

Can also do experiments that produce even more quantitative data. - As the number of items increases in conjunction, the time should increase. Whereas feature stays the same Limits of Behavior When we get detailed behavioral data, we get a bunch of independent measures of the same behavior - Thus run hundreds of conjunction and different participants, average the trials within people and then average all people. Compare the size of our effects to the variability of people. Noise: - If the size of the effect exceeds the amount of variability by a certain factor then that is a significant effect. - Have to reduce the number of random sources of variability as much as possible Sampling: - We generalize behavior, but can’t test everyone - Thus test people available. Can create bias and have to be mindful as to how far you can generalize your data Low Dimensional Data - We have been studying behavior much longer than we have studied neurons. “Why bother studying neural correlates if you can just study behavioral” . - Behavioral data is indirect relative to the cognitive mechanism you are trying to make inferences about. - Behavioral data is a measure of the output of the process. Not of the process - Even if you come up with really clever manipulations to disentangle this or that component from RT, there will always be a huge number of assumptions - Also the accuracy and RT are low dimensional (restricted), can be faster/slower, more/less. This leaves a l...