Lecture 20 - Cortical Processing and Scene Analysis Flashcards
basic unit
frequencies of sound
auditory cortex is hierarchically organized
Neural signals travel through the
core (including A1), then belt,
followed by the parabelt area.
– Simple sounds (pure tones) cause activation in the core area.
– Belt and parabelt areas are
preferentially activated in response to more complex stimuli (tones, voices) made up of many
frequencies.
Simple sounds
retain tonotopy in the cochlea
cause activation in the core area.
belt and parabelt areas
are preferentially activated in response to more complex stimuli (tones, voices) made up of many
frequencies.
Bendor & Wang (2005)
experiment about tonotopy beyond A1
presented complex harmonic tones with the same fundamental frequency (182 Hz). [Work done in monkeys.]
• Single-cell recordings were done outside A1 (belt and parabelt area)
• Six different sounds were played, each made up of diff harmonics
Bendor & Wang (2005) recorded from a cell just outside A1 (belt and parabelt) and
found that it responded to…
.....complex tones with the same fundamental frequency (fundamental frequency perceived as the same basic tone). Such cells were called pitch neurons
(basically same cell responded to all those complex tones because it like the fundamental frequency even though the frequencies themselves different)
- Demonstration of hierarchical organization in auditory cortex.
- Also links perception to physical quality of sound.
pitch neurons
cells just outside of A1 that respond to complex tones with the same fundamental frequency
right outside of A1
seem to like particular perceived pitches: always looking for that repeating pattern established by the spacing of the two frequencies
organized by HOW WE PERCEIVE the stimulus, not their physical components
Are the representations modular?
Effect of training (experience-dependent plasticity) on tonotopic maps:
– Owl monkeys were trained to discriminate between two frequencies near 2,500 Hz.
– Trained monkeys showed tonotopic maps (in A1) with enlarged areas with neurons that responded to 2,500 Hz compared to untrained monkeys.
Importantly, the monkey’s were better at the task after training.
- the cells spread out: it wasn’t just one small region that responded to the particular frequencey (had more cortex working on the discrimination problem: IT CHANGED!!)
- there is certainly hierachical organization, but not necessary modular
Single cell fires to the fundamental frequency….
….regardless of which harmonics
are played.
we do have a well ordered tonotopic map, BUT it is….
malleable, not purely innante, not static, changes over time!!!
change the environment or stimuli = change the neural representation
Experience-dependent plasticity
Experiment by Fritz et al. (single-cell recordings)
– Ferrets were trained to stop licking a water spout in response to a pure tone embedded within a stream of complex tones.
– Single neuron recorded before and after training. Originally responds best to ~9 kHz tone.
– The neuron became quickly tuned to the target frequency (6 kHz, blue arrow) and maintained the effect for hours after the testing session.
this task was starting to take over some of the functionality of other neurons:
!!! there’s a lot of plasticity, and plasticity is enforced by behaviors and environment !!!
What, or ventral stream
starts in the anterior portion of the core and belt and extends to the prefrontal cortex.
– Responsible for identifying
sounds: being able to tell what pitch a tone is, being able to tell tones or pitches apart
Where, or dorsal stream
starts in the posterior core and belt and extends to the parietal and prefrontal cortices.
– Responsible for locating sounds (auditory localization task).
fMRI evidence for
what/where streams
– Pitch recognition (what)
tasks activate more ventral regions.
– Location detection (where)
tasks associated with greater
activation in dorsal areas.
Brain damage evidence
Patient with temporal lobe damage (‘what’ pathway) has trouble recognizing sounds, but can localize.
- Patient with parietal lobe damage (‘where’ pathway) can recognize sounds, but is very poor with localization.
- Nice double-dissociation!!
Patient with temporal lobe damage
has trouble recognizing sounds, but can localize.
can’t say that that’s a bird or a car horn
can’t tell if the pitch went up or down
Patient with parietal lobe damage
can recognize sounds, but is very poor with localization.
trouble with spatial processing for auditory tones
Double-dissociation of
‘what’ and ‘where’ functions
– Temporarily deactivating the anterior auditory area (‘what’ pathway), impaired a cat’s
ability to distinguish tone
patterns (but spared localization).
– Deactivating the posterior auditory cortex, (‘where’ pathway), eliminated the ability to localize sounds (but spared identification).
anterior auditory area
‘what’ pathway
posterior
auditory cortex
‘where’ pathway
Transduction takes place at the
Organ of Corti
fundamental problem with audition
anytime there is sound coming in you have sound pressure waves coming in from different locations and lots different sounds coming in, and all of that is immediately getting mixed in the cochlea
how do we mix these things apart?
Auditory Scene:
The array of
all sound sources in the
environment.
How do we make sense of all the sounds coming to us?
Auditory Scene Analysis:
Process by which sound sources in the auditory scene are separated into individual perceptions.
– This does not happen at
the cochlea since simultaneous sounds are together in the pattern of vibrations on the basilar
membrane.
The auditory scene: perceptual heuristics
A number of auditory grouping principles help us organize
elements of the auditory scene
( we don’t really know wtf is happening)
1) location
2) onset time
3) Similarity of timbre and pitch:
4) Proximity in time -
5) Auditory Continuity
6) Effect of past experience [familiarity]
location
A single sound source tends to be at one location and to move
continuously. [Common fate?]
if a particular set of sounds is coming from a particular location we group those sounds to that location: those sounds belong to that space
if that sound source starts to move then the sounds start to move with it
Onset time
Sounds that start at different times are likely to come from
different sources, if sounds start together they probably came from the same source
Similarity of timbre and pitch:
Similar sounds are grouped together (e.g. flute trills belong to the flute, while trumpet blares belong to the trumpet.) [Similarity? we group things of like colors or shapes, etc…]
– Strong CUE: Such grouping allows us to perform auditory stream segregation based on attributions of different timbres or pitch to multiple sources.
similarity of PERCEPTUAL properties
auditory stream segregation
identify extremes of notes that are changing together, but they have the same pitch and timbre
segregate complex streams and keep these streams distinct
proximity in time
- sounds that occur in rapid succession usually come
from the same source. [Proximity?]
greater the time lag (based on experience) the more willing you are to break up those sounds
The auditory scene: perceptual heuristics
Compound melodic line in music is an example of
auditory stream segregation
Experiment by Bregman and Campbell (1971) demonstrated how compound melodic line in music is an example of auditory stream segregation.
– Stimuli were alternating high and low tones
– When stimuli played slowly, the perception is one of hearing high
and low tones alternating (not broken into diff melodic lines - don’t hear separately).
– When the stimuli are played quickly, the listener hears two streams; one high and one low (hear them at the same time).
– Pulled together by the similarity of pitch and timbre groups the notes, but only at high
speeds (only if the proximity allows it).
Grouping by similarity experiment by Bregman & Rudnicky (1975)
– Listener hears two standard tones (X & Y) and can easily perceive the order.
– Two distractor (D) tones are placed around X & Y and the listener now has difficulty perceiving the order.
– Adding a series of captor tones (C) forms a stream with the D tones, and the listener can once again easily hear the order of X & Y.
Outer hair cells act as a ____ and heighten sensitivity of inner hair cells to a _______ .
cochlear amplifier
characteristic frequency.
Place Coding
suggests that the specific locations on the basilar membrane are stimulated by certain frequencies.
This results in a tonotopic map
tonotopic map
high frequencies activating hair cells near the base and low frequencies active at the apex.
Signals are sent from the cochlea via a sub-cortical route
[CoNuc SON-IC MG] to the
primary auditory receiving area
in temporal cortex [A1].
In A1 we have tonotopic organization that becomes
increasing….
Signals then go to ____and ____
streams for further processing.
…..hierarchical in core, belt and parabelt regions.
dorsal (Where?) , ventral (What?)
Grouping by pitch similarity
A repeating tone (pitch) is able
to capture an ascending and
descending scale as they cross.
once pitches become close enough
it’s hard to distinguish them
sounds become tangled and are sort of “galloping”
auditory continuity
Sounds that stay constant or change smoothly are usually from the same source. [Good continuation?]
we’re explicitly tuned to hearing gaps
Auditory continuity experiment by Richard Warren et al. (1972)
Tones were presented and interrupted by gaps of silence or noise.
– In the silence condition, listeners perceived that the sound stopped during the gaps.
– In the noise condition, the perception was that the sound continued behind the noise.
things that are continuous (smooth background) will fade and the thing that grabs our attention will come to the foreground
Effect of past experience [familiarity]
Experiment by Dowling & Harwood (1986)
• Melody “Three Blind Mice” is played with notes alternating
between octaves
• Listeners find it difficult to identify the song
• But after they hear the normal melody, they can then ‘hear’
(recognize) it in the modified version using melody schema.