Transcript Chapter 10 聲音，聽覺系統與音調知覺

Chapter 11
聲音，聽覺系統與
音調知覺
聽覺可以感知視覺系統某些不及之處
2
聽覺之功能
• 訊號（signaling）
– 過街時的導盲鈴、身後的腳步聲
• 溝通（communication）
– speech
– Blindness isolates you from things, but deafness
isolates you from people.
Helen Keller
• 樂趣（pleasure）
– 音樂
什麼是聲音？--物理或知覺屬性
• 森林中一棵大樹倒下，但沒有人聽到，
這樣算有聲音嗎？
4
– Yes,
聲音來自空氣或其他介質中的壓力變動
– No,
聲音是來自於「聽」的經驗
聲波
5
聲音的物理與知覺屬性
•聲波
–速率340m/sec
(light 1,500m/sec)
•振幅（Amplitude）
振幅愈大，響度愈大
Fig. 11-3, p. 236
7
8
– 振幅測量
• 反映最大與最小的差距，但壓縮較大值端，以模擬心
理響度
• 分貝（dB）- after A. G. Bell
• 20 log (p/p0)
– p-聲源聲壓
– p0 - 比較基準（1000Hz純音恰可被聽到的閾值水準，約為
20 micropascals）
• eg., p=20
20 log (20/20)=20 x 0=0 dBSPL
100倍
+40 dB
p=2000（100倍）
20 log (2000/20) = 20 x 2=40 dB SPL
10倍 →+20 dB
100倍→ +40 dB
1000倍→ +60 dB
響度隨強度（dB）呈線性增加
10
– 頻率（Frequency）與音調（pitch）有關
•Hertz-cycles/sec
500 Hz
1000 Hz
4000 Hz
11
Fig. 11-5, p. 237
Complex Tones
• The repetition rate of a
complex tone :
Fundamental Frequency
= First Harmonic
Ex. 200 Hz
First harmonic
Second harmonic
Third harmonic
Fourth harmonic
12
• Removal of the first
harmonic results in
a sound with the
same perceived
pitch, but affect
perception of the
tone.
• The repetition rate is
still 200 Hz.
13
八度（octave）
• A0=27.5, A1=2xA0, A2=4xA0,………..
– 20~20000 Hz為
人類可聽範圍
(range of hearing)
– 可聽曲線
（audiblity curve）
• 最敏感20004000 Hz—與
speech有關
聽覺反應區（auditory response area）
15
• Threshold for feeling
– OSHA 不可大於90 dB (8hr/day)
– 其他動物
•更低頻：大象，鴿子
•更高頻：海豚，狗
16
• 響度由聲壓水準及頻率共同決定
– eg., 40dB, 100Hz -- B
40dB, 1000Hz --C
– B與C並未落在同一條等響曲線上—響度不
同（C>B）
• 等響曲線（equal loudness curve）
17
– eg., 與40dB1000Hz相等響度的音構成“40”
曲線
– 音強時，各種頻率的感受性相近；但音弱時，
最高或最低頻音都不易聽見
•「LOUDNESS」鍵
18
• Perceiving sound: pitch, loudness, and
timpre.
• 音色（Timbre）
– 純音vs.複合音
•單一vs. 多種頻率組成
– White noise
– Complex sound
•樂器的差異
– 樂音多由若干頻率純音組合而成，但其他
日常生活中常見的聲音則更為複雜
19
純音440Hz
2nd harmonic 880Hz
3rd harmonic 1320Hz
– 相加合成（additive
synthesis）
• 由fundamental frequency
加上harmonics（是
fundamental frequency 的
倍數）
• 組合方式界定每種樂音
的獨特性
20
Fig. 11-8, p. 240
Frequency spectrum
21
22
Fig. 11-10, p. 240
聽覺系統的結構與功能
• 聽覺系統需要達成三項功能
– 將聲波傳達至受器
– 將聲波所傳達的氣壓變動轉換為電訊號
– 設法使電訊號傳遞如音調，響度，音色，
定位等聲音的屬性
23
• 聲音如何抵達受器？
• 外耳（outer ear）
– Pinna
– Auditory canal
•約3cm，保護耳膜，中耳
•對2000-5000 Hz有放大效果（resonant
frequency of the canal）
24
Fig. 11-11, p. 242
• 中耳（middle ear）
– 耳膜（tympanic membrane）
– 聽小骨（ossicles）
Tympanic membrane
 malleus（槌骨）
 incus（砧骨）
 stapes（鐙骨）
 oval window
– 為何需要三塊聽小骨？
• 由外耳/中耳低密度的空氣至內耳較高密度的液體時，聲波
的震動只能傳遞極少的部分
• 聽小骨協助放大訊號以解決這個問題
– 魚就不需要「中耳」的功能
27
– 如何放大？
增
加
22
倍
• 將較大面積的耳膜振動集
中於較小面積的鐙骨以增
加單位面積接收的壓力
• 聽小骨的槓桿運作
– Acoustic reflex
• Middle-ear muscles 與聽小
骨連結，在音量極高時收
縮來牽制聽小骨的動作
28
• 鐙骨振擊卵形窗造成內耳
（inner ear）液體振動
• 內耳結構
– 耳蝸（cochlea）
• 充滿液體
• scala vestibuli（上半）
• scala tympani (下半)
• Cochlea partition
– organ of Corti
有hair cell—聽覺受器
Inner hair cell vs. outer hair cell
耳蝸內液體振動造成cochlea
partition 上下運動，hair cell頂端
纖毛彎曲，傳送電訊號
– basilar membrane
– tectorial membrane
30
stapes oval window liquid in scala vestibuli basilar
membrane
stapes pulls back, then basilar membrane 
Organ of Corti 
Tectorial membrane 
31
• Cilia bend in a specific
direction  Ion channels
opening  Ions flow
across the cell membrane
Electrical signals
The release of neural
transmitter from the inner
hair cell
32
– Inner hair cell 之纖毛彎曲
造成depolarize 或
hyperpolarize （釋放神經
傳導物質或停止傳導），
傳送電訊號
– 可以產生電訊號的纖毛彎
曲幅度不大—敏感度高
33
Place theory of hearing
• 聽覺系統如何代表頻率訊息？
34
• Bekesy’s theory
(1961諾貝爾生理與醫學獎得主）
– traveling wave motion of the basilar membrane
35
•Basilar membrane的基部（base）較窄，較緊
36
•Envelope of the traveling wave可以顯示不同部位
hair cell受到最大影響的幅度
•P點附近的hair cell送出最強的神經訊號
•P點的位置受聲音頻率的影響
– P點的位置受聲音頻
率的影響
• 低頻-apex
高頻-base
38
– 位置編碼的生理證據
• tonotopic map
– 低頻.-apex
高頻.-base
39
•測量恰可引發神經元反應的dB SPL，可得頻
率調適曲線（tuning curves）
– 該神經元最為敏感的頻率稱為characteristic frequency
40
– 貓聽神經的 frequency tuning curve
41
• 心理物理證據
– auditory masking
pp. 275
Envelope的重疊程度解釋何以有不對稱的遮蔽效應
43
• tuning curve 為何很窄？
– Outer hair cells產生的運動影響basilar membrane
的運動
• 依頻率使特定範圍的basilar membrane 活動受強化
高頻音—base
低頻音--apex
44
• Basilar membrane上的頻率分析
– Cochlea包含多組濾波器（filters），
每組處理特定的頻率範圍
45
46
Fig. 11-30, p. 251
– 一系列的tuning curve，
低點連起來恰符合
audibility curve 的型態
• 顯式有一系列的頻率分
析器，各自負責很窄的
範圍，而產生整體的可
聽曲線
• 也符合聽覺神經元的
tuning curve
47
Basilar membrane對複合音的反應
複合音的不同組成造成對不
同頻率反應的f濾波器產生
反應
Basilar membrane 對不同的頻率組成
產生反應
--在進行Fourier analysis ?
Updating Békésy’s Place Theory
• Distinguishing close
frequencies ?
• New research
– Using “live membranes”
shows that there is less
overlap between nearby
frequencies.
• Both cell elongating and
contracting increase the
motion of the basilar
membrane and sharpen its
response to specific
frequencies.
The action of the outer hair cells:
cochlear amplifier
Figure 11.32 Effect of OHC damage on frequency tuning curve. The solid curve is the frequency
tuning curve of a neuron with a characteristic frequency of about 8,000 Hz. The dashed curve is the
tuning curve for the same neuron after the outer hair cells were destroyed by injection of a chemical
(Adapted from Fettipalce & Hackney, 2006).
Timing theory of frequency coding
• 時間編碼（timing coding)
– Frequency of the stimulus is signaled by the
frequency of nerve firing?
• does not work because the max. nerve firing rate = 500
impulses/second
– phase-locking
• Neurons fire in synchrony with the phase of a stimulus:
– Fire only at peaks
• Larger number of fibers fire in response to high
frequency than low frequency
51
52
Fig. 11-34, p. 252
Fig. 11-35, p. 253
Hearing Loss
• Two types
– Conductive hearing loss
• Blockage of sound from the receptor cells
– Sensorineural hearing loss
• Damage to hair cells, the auditory nerve, or brain
• Most common type: Prebycusis
Hearing Loss
• Presbycusis
– Greatest loss is at high
frequencies
– Affects males more
severely than females
– Apparently caused by
exposure to damaging
noises or drugs because
people in preindustrial
cultures often do not
experience large
decreases in highfrequency hearing in old
age
Hearing Loss
• Noise-induced hearing loss
– Loud noise can severely damage the hair cells
– MP3 players: “Leisure noise” can also cause hearing
loss
3-hour Game
Safe Level
57
聽皮質的頻率分析
• 聽覺通道（auditory
pathway ）
• cochlear auditory nerve
fiber  cochlear nucleus
 superior olivary
nucleus (brain stem) 
inferior colliculus
(midbrain)  medial
geniculate nucleus
(thalamus, near LGN) 
primary auditory
receiving area (A1, in
temporal lobe)
58
• 階層處理
– (in monkey) core (simple sounds) →belt
(complex sounds, eg., monkey
calls)→parabelt
與視覺消息處理性質相近
59
Fig. 11-37, p. 254
• 聽覺的what & where
– What路徑始於core與belt的前區，至側葉，前額葉
– Where路徑始於core與belt後區，至頂葉與前額葉
60
Fig. 11-38, p. 255
– 神經生理證據
•JG側葉受損，
無法從事聲音辨
認
•ES頂葉，前額
葉受損，無法從
事聲音定位
61
Fig. 11-39, p. 255
•What task – 辨認音調（pitch）
where task – 偵測位置
62
Fig. 11-40, p. 255
• A1在音調知覺中的角色
– Tonotopic map
63
– 訓練增加owl monkey A1對於相關頻率的反
應區域面積
Fig. 11-42, p. 256
– 聽覺皮質受損的pt
65
Fig. 11-43, p. 257
• Bendor and Wang (2005): Pitch neurons in the
marmoset auditory cortex
– These stimuli with different frequencies were perceived as
having a pitch corresponding to the 182-Hz fundamental
frequency.
– The corresponding cortical neurons responding only to stimuli
associated with the 182-Hz tone
Figure 11.43 Records from a pitch neuron recorded from the marmoset auditory cortex. (a) Frequency spectra for
tones with fundamental frequency of 182 Hz. Each tone contains three harmonic components of the 182 Hz
fundamental frequency; (b) Response of the neuron to each stimulus. (Adapted from Bendor & Wang, 2005). 66
• Experience dependent plasticity
– human hearing
• 25% more auditory cortex was activated by
piano tones in musician vs. nonmusician (and
the electrical activity was twice as strong in
musician)
– Task related plasticity
• Neuron in a marmoset’s auditory cortex
• Training
– complex sound → lick water
pure tone (of a particular freq) → stop licking
– After a few trials, the neuron’s response profile
has changed
• The auditory sys tem shapes its neurons to
behaviorally important stimuli
67
Green – average level of firing
Blue – decreased firing
Red/yellow –increased firing
68
Figure 4.24 (a) Greeble stimuli used by Gauthier. Participants were trained to
name each different Greeble. (b) Brain responses to Greebles and faces
before and after Greeble training. (a: From Figure 1a, p. 569, from Gauthier, I.,
Tarr, M. J., Anderson, A. W., Skudlarski, P. L., & Gore, J. C. (1999). Activation
of the middle fusiform “face area” increases with experience in recognizing
novel objects. Nature Neuroscience, 2, 568-573.)
69
70
Shepard tone (scale)
71
Cochlear Implants
• Electrodes are inserted into the cochlea to
electrically stimulate auditory nerve fibers.
• The device is made up of:
– a microphone worn behind the ear,
– a sound processor,
– a transmitter mounted on the
mastoid bone,
– and a receiver surgically mounted
on the mastoid bone.
Figure 11.46 Cochlear implant device. See text for details.
Cochlear Implants
• Implants stimulate the cochlea at different
places on the tonotopic map according to
specific frequencies in the stimulus.
• These devices help deaf people to hear
some sounds and to understand
language.
• They work best for people who receive
them early in life or for those who have
lost their hearing, although they have
caused some controversy in the deaf
community.