Unformatted text preview: From Ear to Brain Cognitive Neuroscience The computational problems n Identify what the sounds are and where they are coming from q What: n Object constancy (invariance) Sounds and words must be recognized across variability in: noise source (speaker) q Where: n Location must be computed because (unlike vision) it is not "given" by the geometry of the sensory surface How do we convert air pressure into neural activity and perception? Sound waves moves the tympanic membrane > Tympanic membrane moves the ossicles > Ossicles move the membrane at the oval window > Motion at the oval window moves the fluid in the cochlea > Movement of the fluid activates sensory neurons Why so many steps? n Problem: Fluid in the inner ear doesn't move as easily as air in the outer ear q Need to amplify the pressure at the oval window so that it is 20x what it is at the tympanic membrane q n n Ossicles transform tympanic membrane movements into smaller but stronger movements of the oval window The surface area of the oval window is smaller than the surface area of the tympanic membrane Basilar membrane: Narrower and less flexible at the base, broader and more flexible at the apex Differentially deformed by sound waves of different frequencies (as frequency decreases, maximal deformation moves towards apex) Mechanical energy must be converted to membrane polarization n n n Organ of Corti: hair cells, support cells Hair cells: auditory receptors Stereocilia of hair cells: are bent thru the movements of the basilar membrane n The movement causes: q Opening and closing of K+ channels in stereocilia Activates voltagegated Ca++ channels Causes neurotransmitter release from base of hair cell onto dendrites of: q q n Spiral ganglion cells q Their axons form the auditory nerve (cranial nerve VIII) The first cells in the auditory pathway to generate action potentials q Auditory Pathway n All nuclei above the cochlear nuclei receive input from both ears Considerable processing prior to cortex n Primary Auditory Pathway Characteristics of sounds Intensity (amplitude) n Frequency n Location
n Encoding Intensity n 1. Intensity, coded by means of: Firing rate: q q q Greater amplitude of vibration of basilar membrane Greater polarization of hair cells Greater number of action potentials in spiral ganglion cells Greater amplitude of vibration of basilar membrane Larger area of basilar membrane affected Greater number of hair cells releasing neurotransmitters 2. Number of active neurons: q q q Encoding Frequency 1. Frequency tuned neurons: q Tonotopic "maps": nearby neurons have similar characteristic frequencies (in auditory nerve, relay nuclei, the MGN, auditory cortex) Encoding Frequency q Frequency tuned neurons can't be the only source of frequency info cause: n n At a fixed frequency, a more intense sound will produce a maximal deformation at a point further up the basilar membrane than a less intense sound There aren't neurons with low characteristic frequencies (below 200 Hz) Encoding Frequency 2. Phase locking: n Action potentials are produced in phase with sound wave q q Potentials fire at peaks, troughs or some other constant A group of neurons that fire on the cycles together can represent sound frequency (volley principle) n However, beyond 4k HZ, the variability in timing of action potentials is comparable or greater than the time intervals between successive cyclesphase locking not a good mechanism for encoding high frequency info Encoding Frequency n Mechanisms of frequency encoding: Low freq q Intermediate freq q High freq q phase locking phase locking and freq tuning frequency tuning Auditory Cortex: Tonotopy (tonotopic map) Primary auditory cortex (A1) is on the anterior tranverse gyrus of the superior temporal gyrus. Encoding Horizontal Location: Interaural Time Differences (ITD) Speed of sound 340.29 m/sec
n For a sudden, discrete sound the interaural time delay is 0 0.6 ms (assuming a 20 cm head) n For a continuous tone, ITD affects the arrival time of peaks q q 200 Hz tone: one cycle = 172 cm 20,000 Hz one: one cycle =1.72 cm n ITD not useful for frequencies were one cycle is smaller than the distance between the ears (> ~2000 Hz) Interaural Intensity Differences n The head casts a sound shadow differences in the intensity of the sound received by each ear. However, for low frequency sounds the cycles are so large that the shadow has little effect. n Encoding Location n Duplex theory of sound localization: 202000 Hz q 200020,000 Hz q ITD IID n How do neurons detect ITDs? Brain stem neurons show preferred ITDs q How is this info computed?
q A circuit for detection of interaural time differences in the brain stem of the barn owl (Carr and Konishi 1990) n Barn owls can capture prey using only auditory cues. Can make fine distinctions between frequencies to distinguish potential prey from wind, grass, nonprey. Facial ruff: n n n Amplifes sound n Ears are set asymmetrically! Computing ITD: Delay lines and coincidence detection Proposed in 1948.
n The different neurons will be maximally responsive to specific differences in the times at which sounds arrive to the right and left In this way each neuron each neuron has its preferred ITD and encodes the location of the sound source n Computing ITD Detects ITDs of < 0.005 ms !!! (Carr & Konishi, 1990) Topdown influences: Categorical Perception n da n ga Present sound tokens that are equidistant Listeners perceive a categorical boundary Speech > silence Areas that are sensitive to phonological (vs. merely acoustic) differences Shestakova et al. 2004 n n 150 exemplars each of [a], [i], [u] from 450 speakers MEG: 11 subjects, passive listening Source Localization of M100
c) Representative subject d) Group data: Significant difference in zcoordinate between [a] and [i], in both hemispheres. Evidence for phonemotopic map. More complex response properties Binder et al. (2000) ...
View Full Document
- Spring '08
- Action Potentials, hair cell, basilar membrane, oval window, tympanic membrane