Language experience shapes representation of pitch relevant neural activity in the human brainstem

Ananthanarayan (Ravi) Krishnan

Research supported by NIH Grant, 1R01DC008549

Auditory Electrophysiology Laboratory

FFR Workshop, May22-23 UCL, London

Why study Pitch?

Ø Pitch  plays  an  important  role  in  the  percep1on  of  speech,  language,  and  music.  

Ø Therefore  there  is  considerable  interest  in  understanding:  Ø   (i)    How  pitch  relevant  informa2on  is  representa2on  at  both                      subcor2cal  &  cor2cal  levels;    Ø (ii)    How  long  term  language  experience  influences  this                        experience-­‐dependent;  and  

Ø (iii)  The  nature  of  interplay  between  cor2cal  and                        subcor2cal  components  in  the  pitch  processing                      hierarchy  

Pitch processing-hierarchical-includes both cortical and sub-cortical stages of processing

Language experience shapes pitch processing

Pitch processing of lexical tones at the cortical level is influenced by language experience.

4

Present results from the brainstem that supports the view that language experience also influences representation of pitch relevant neural activity in the brainstem

Experimental Strategy: Cross-linguistic approach  

 Examine: 1. native & non-native contours

2. context (speech vs. non-speech) of contours 3. sensitivity to temporal regularity 4. relationship to behavioral measures 5. relationship between brainstem, cortical and

behavioral measures relevant to pitch Subjects: Chinese (Mandarin) & English subjects:

right handed, no music experience Mean age when Chinese started to learn English: 11 yrs Dependent Variables:

Pitch Tracking Accuracy Pitch strength (periodicity strength)

Methods:  Chinese  Subjects  

Ø Normal  hearing    na-ve  speakers  of  Mandarin  Chinese  Ø   Closely  matched  in  age  (24.50  ±  3.53  years),  years  of  formal  educa1on  (16.90  ±  2.88  years),  and  were  strongly  right  handed  (laterality  index  93.10  ±  11.22%).    

Ø All  par-cipants  were  born  and  raised  in  mainland  China.    

Ø None  had  received  formal  instruc1on  in  English  before  the  age  of  nine  (12  ±  1.24  years).  

Ø LiIle  or  no  music  experience      

Krishnan et al. Hearing Research (2004); Cog. Br. Res (2005)

Clothing

aunt

chair

easy

Stimuli: Lexical Tones

Cross-linguistic approach: Native Mandarin vs non-native English Listeners Stimuli: lexical tones Note: Identical spectra with only pitch contour changing

1. Is brainstem neural activity relevant to pitch influenced by language experience?

3

Mandarin Tone 2 PITCH TRACKING

PERIODICITY STRENGTH

Encoding of pitch is sensitive to language experience

Krishnan, Xu, Gandour, & Cariani Cognitive Brain Research 2005

PT: crosscorrelation between Fo contours extracted from stimuli and FFRs PT: crosscorrelation between Fo contours extracted from stimuli and FFRs PT: crosscorrelation between Fo contours extracted from stimuli and FFRs PS: magnitude of normalized autocorrelation peak per time frame

Blue stripe shows the dominant interval for rising tone; 10 ms = 100 Hz; period is the inverse of frequency

Krishnan et al. Cognitive Brain Research 2005

Chinese English

More coherent phase-locking to F0 interval in the Chinese

FFRs preserved the pitch-relevant information of lexical tones in both native and non-native listeners

Summary

Stronger representation of in native listeners in terms of both tracking accuracy and periodicity strength

These results suggest long term language experience induced neural plasticity at the brainstem level may be enhancing or priming linguistically-relevant features of the speech input

3

Xu, Krishnan, & Gandour NeuroReport 2006; Krishnan et al., (2009)

Will linear & curvilinear variants of a higher degree polynomial still yield such results?

2. Is this language experience-dependent plasticity sensitive to shape of pitch contours?

Stimuli used by Krishnan et al. (2005) – curvilinear f0 contours that were modeled after Mandarin tones in natural speech.

If brainstem reorganization is induced by language experience, will deviations from prototypical f0 contour still produce an experience-dependent effect?

3

Xu, Krishnan, & Gandour NeuroReport 2006

Sensitivity to shape of pitch contours: linear and non-linear contours in speech context

•  No crosslanguage differences in periodicity strength or tracking accuracy for linear contours

3

LINEAR

CURVILINEAR

T2L

T4

T4L

T2 •  Representation of pitch-relevant information appears to dependent on specific fine-grained features of pitch contours that occur in natural speech

•  PT & PS: Chinese > English for T2 only •  English = Chinese for curvilinear (T2i) and linear (T2 l, T2tl) variants of T2

Time (sec) 0.005 0.045 0.085 0.125 0.165 0.205 0.245

Freq

uenc

y (H

z)

100

105

110

115

120

125

130

135

Sensitivity to shape of pitch contours: linear and non-linear contours in a nonspeech context

Krishnan, Gandour, Bidelman, & Swaminathan NeuroReport (2009)

T2

T2i

T2l

T2tl

Summary

Brainstem representation of pitch relevant information in the Chinese appear to be acutely sensitive to dynamic changes in trajectory throughout the duration of a pitch contour

Conclude that long-term experience-dependent influence on brainstem representation of pitch relevant information is specific to contours that are part of their experience.

Conclusions 3. Is this language experience-dependent plasticity specific to speech context ?

Typically used non speech stimuli (rotated, sine wave, speech in noise etc) has some speech characteristics to it.

To answer this question the pitch contours have to presented in a non-speech context

We used dynamic curvilinear iterated rippled noise (IRN) stimuli to remove any potential lexical-semantic confound

3

15 Input, x(t)

Pitch  contours  in  a  non-­‐speech  context:  IRN  s1muli  

T ime (sec )

0 0.25

Fund

amen

tal f

requ

ency

(Hz)

80

150

80

90

100

1 1 0

120

130

140

150

0 0.05 0.10 0.15 0.20 0.25

NON-SPEECH (IRN; N=32)

T1

T2

T3

T4

SPEECH

IRN block diagram

Delay d

Gain g

+ Delay d

Gain g +

Output, y(t)

Swaminathan et al, IEEE (2008)

D

T1  

Krishnan, Swaminathan, & Gandour Journal of Cognitive Neuroscience (2008)

FFR: Mandarin tone; nonspeech; curvilinear; native/nonnative

1  

T2  

T3  

T4  

PITCH  STRENGTH  

Periodicity Strength (segmental analysis)

2 Swaminathan, Krishnan & Gandour Neuroreport (2008)

C > E on linguistic tones for both S & NS- particularly in segments showing rapid changes in pitch

18

Are language effects dependent on context? PITCH STRENGTH

C > E on linguistic tones both S & NS contexts and particularly in sections with rapid pitch changes

Swaminathan et al, NeuroReport (2008)

For both groups, pitch strength for speech greater than for non-speech

Pitch Strength: Speech > Non-Speech

In native listeners, representation of dynamic IRN stimuli were superior in terms of both tracking accuracy and pitch strength

3

These findings suggest that the experience-dependent enhancement of pitch relevant information is not specific to speech domain; and shows particular sensitivity to the dynamic portions of native pitch contours.

Summary

Segmental evaluation suggested better sensitivity in the native listeners to rate of change of pitch

Iteration steps (n) n  =  4  n  =  2   n  =  8   n  =  12   n  =  16   n  =  32  

PITCH  CONTOUR   PITCH  STRENGTH  

4. Is this experience-dependent enhancement in pitch representation sensitive to changes in temporal regularity?

Krishnan, Bidelman, & Gandour Hearing Research (Brain Res., 2009; Hear Res., 2010;) 6

Pitch  relevant  informa1on  appears  to  emerge  at  a  lower  itera1on  number  in  Chinese  listeners  

Pitch tracking accuracy better in Chinese and at a smaller iteration step

Pitch Strength grows more rapidly with iteration steps for Chinese listeners

Summary

Pitch-tracking accuracy & periodicity strength shows greater sensitivity to pitch salience in the native tone language group (Chinese)

Periodicity strength emerges 2–3 times faster in the Chinese group with increasing pitch salience

These results suggest that long term language experience enhances the representation of pitch relevant neural activity in the brainstem that may underlie pitch salience.

Information related to pitch salience emerges early along the auditory pathway in pre-attentive, sensory-level processing

Krishnan, Bidelman, & Gandour Hearing Research (2010) 2

Relationship between neural and behavioral measures

Rela-onship  between  brainstem,  cor-cal,  and  behavioral  measures  relevant  to  pitch  

The  pitch  Onset  Response  (POR)  

Ø Human  POR,  as  measured  by  MEG,  reflects  synchronized  cor1cal  neural  ac1vity  specific  to  pitch  

 Ø POR  latency  and  magnitude,  

for  example,  has  been  shown  to  depend  on  pitch  salience.  

   Ø Anatomical  Source:  Lateral  

Heschl’s  Gyrus-­‐the  puta1ve  site  of  pitch  processing      

!"#"$%"&'("$)*!") +,-.+ !"/"$0"& %1$% %12) "33"4% 5$) 126107 )26'8294$8% :!! ;<;;;=> ?$@0") A $8& =B< C8 D!&"! %D E*$8%237 %1" "33"4%F %1"!"0$%2D8)12# @"%5""8 %1" ("$8 G-H $(#02%*&" $8& %1" 0D6$!2%1( D38D2)" &*!$%2D8 5$) 9%%"& 52%1 $ 028"$! 3*84%2D8 )1D58 28 I26< J$< ?1"&$%$ 5"!" @")% $44D*8%"& 3D! @7 %1" 3*84%2D8

"#$ ! "J!K# 0D6$%&'% " AA!A8+!(

51"!" "#$ 2) G-H $(#02%*&" $8& %&' 2) 8D2)" &*!$%2D8< ?1*)F "/"!7&D*@0286 D3 %1" 8D2)" &*!$%2D8 5$) $))D42$%"& 52%1 $8 284!"$)" 28 G-H$(#02%*&" D3& J<K 8+<(< ?1" 9%%"& 3*84%2D8 $44D*8%"& 3D! KLM D3 %1"/$!2$84" 28 %1" G-H $(#02%*&" &$%$<?1" +,-.+ 3D! %1" ,A;;( &"N"4%2D8 28 %1" ,A;;('@$)"& )D*!4"

5$/"3D!() :8D% 6!$#124$007 &2)#0$7"&B $0)D !"/"$0"& )268294$8% $(#02'%*&" @*% 8D% 0$%"847 "33"4%) :?$@0") A $8& =B< C8 %1" L;;;'() 4D8&2%2D8%1" $(#02%*&" D3 %12) 4D(#D8"8% :"A;<O 8+<(B 5$) )0261%07 !"&*4"& $)4D(#$!"& %D %1" D%1"! %1!"" 4D8&2%2D8) :J;;()F"AP<Q 8+<(> A;;;()F"AP<R 8+<(> =;;;()F "AL 8+<(B< ?1" "33"4% !"$41"& )%$%2)%24$0 )26'8294$84" 28 %1" +,-.+ :?$@0" ABF @*% 8D% 28 %1" )*@)"E*"8% $()* +(,4D(#$!2)D8) :?$@0" =B< ?1" !")*0% 2) 200*)%!$%"& 28 I26< J@<

!"#$%##"&'

S" 1$/" )1D58 %1$% %1" G-H $(#02%*&" 284!"$)") 52%1 %1" %2("28%"!/$0 @"%5""8 %1" D8)"% D3 )%2(*0*) "8"!67 $8& %1" D8)"% D3 #2%41$% %1" %!$8)2%2D8 3!D( 8D2)" %D HCT< ?1" 028"$! 284!"$)" 28 G-H$(#02%*&" 52%1 %1" 0D6$!2%1( D3 %1" 8D2)" &*!$%2D8 !")"(@0") %1"

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

!"#$%&"' "( %)* +,-

Y!239%1) $8& 4D00"$6*") 1$/" 4D8%!$)%"& %1" $4%2/$%2D8 #!D&*4"& @78D2)" $8& $ HCT 28 $ 3*84%2D8$0 @!$28 2($6286 )%*&7 $8& &"(D8)%!$%"& $12"!$!417 D3 #2%41'!"0$%"& $4%2/$%2D8 28 %1" 4D410"$! 8*40"*)F %1"283"!2D! 4D0024*0*)F %1" ("&2$0 6"824*0$%" @D&7 :Y!239%1) -* ".<F=;;AB $8& $ 8D8#!2($!7 4D!%24$0 $!"$ 28 %1" 0$%"!$0 #$!% D3 ZY :Y!239%1)-* ".<F AKKO> Y!239%1) -* ".<F =;;AB< ?1" 28/D0/"("8% D3 0$%"!$0 ZY 28 %1"#!D4"))286 D3 %"(#D!$0 !"6*0$!2%7 5$) 4D!!D@D!$%"& 28 $ ($68"%D"8'4"#1$0D6!$#124 )%*&7 @7 Y*%)41$0[ -* ".< :=;;=B> $ #2%41')"8)2%2/"6"8"!$%D! #!D&*4") $ )*)%$28"& 9"0& 28 %12) !"62D8< C8 D*! !"4"8% )%*&7:\!*(@1D0] -* ".<F =;;PBF %1" 6"8"!$%D!) D3 %1" G-H5"!" 0D4$02]"& %D $#D28% 8"$! %1" ("&2$0 #$!% D3 ZYF $8& %12) )D*!4" 0D4$%2D8 5$)4D89!("& 28 $ )*@)"E*"8% #2%41 )%*&7 @7 H*## -* ".< :=;;PBF 51"!")2(20$! )%2(*02F 4D8)2)%286 D3 8D2)" $8& HCT )"6("8%)F 5"!" *)"&<

.)* +,- $'/ %)* 01223

?1" ,A;;:(B^G=;;:(B'4D(#0"_ 4$8 @" "/D["& @7 $0(D)% $87 [28& D3)*#!$'%1!")1D0& $*&2%D!7 )%2(*0*)> 1"84" %1" &")268$%2D8 X-H< ?1",A;;:(B 1$) @""8 3D*8& %D 4D8)2)% D3 )"/"!$0 )2(*0%$8"D*)07 $4%2/")*@4D(#D8"8%) :.$*61$8 U H2%%"!F AKQ;> SD0#$5 U G"8!7F AKQJ>`4a$00*( U a*!!7F AKO;> ,$b$b%$b8"8 U G24%D8F AKOQ> SDD&)F AKKJ>W*&& -* ".<F AKKOBF $8& %1" 4"8%!" D3 $4%2/2%7 2) 28 %1" !"62D8 D3 #0$8*(%"(#D!$0" :V*b%["81Db8"! U T%"28)%!$b%"!F AKKOB< +) %1" $(#02%*&" $8&0$%"847 D3 %1" ,A;;:(B $!" 0$!6"07 &"%"!(28"& @7 %1" 28%"8)2%7 :H$#28-* ".<F AKRR> W"$60"7 U \8261%F AKRQB $8& !2)"'%2(" D3 %1" )D*8&:W2"!($88 U Z"20F =;;;BF %1" ,A;;:(B 1$) %!$&2%2D8$007 @""8 !"6$!&"&$) $8 "8"!67'D8)"% !")#D8)"< ZD5"/"!F )7)%"($%24 28/")%26$%2D8) 1$/"!"/"$0"& %1$% %1" ,A;;:(B 2) $0)D )"8)2%2/" %D )%2(*0*) 3"$%*!") D%1"!%1$8 "8"!67F )*41 $) )#"4%!$0 4D(#D)2%2D8 :HD@"!%) -* ".<F =;;;>V*b%["81Db8"! -* ".<F =;;A> T"2%1"!'G!"2)0"! -* ".<F =;;PB $8& #"!2D&242%7:H$6D% U V"#$*0'X!4D0"F AKKR> T"2%1"!'G!"2)0"! -* ".<F =;;=> V*b%["81'Db8"! -* ".<F =;;P@B< W$)"& D8 ("$)*!"("8%) D3 )%2(*0*) 28%"6!$%2D8%2("F +0$28 -* ".< :AKKQB )*66")%"& %1$% %1"!" $!" $% 0"$)% %5D 3*84%2D8$0)*@)7)%"() *8&"!07286 %1" 6"8"!$%2D8 D3 %1" ,A;;:(Bc D8" 5124128%"6!$%") )D*8& "8"!67 52%1 $ 4D(#$!$%2/"07 )1D!% %2(" 4D8)%$8% $8&$8D%1"! 51241 2) 28/D0/"& 28 $8$07)286 )D*8& 3"$%*!")F )*41 $) #2%41F%2(@!" D! &*!$%2D8F 52%1 $ 0D86"! %2(" 4D8)%$8%<?1" 4D!!")#D8&"84" @"%5""8 %1" #D0$!2%7 D3 %1" ,A;;( $8& %1$% D3

%1" G-H )*66")%) %1$% G-H $(#02%*&" 2) !"0$%"& %D %1" ,A;;(F @*%%1"!" $!" $0)D $!6*("8%) $6$28)% %12) 28%"!#!"%$%2D8< I2!)%07F %1" ,A;;($8& %1" G-H 1$/" ($!["&07 &233"!"8% 0$%"842")< T"4D8&07F \!*(@1D0]-* ".< :=;;PB )1D5"& %1$% %1"2! )D*!4" 0D4$%2D8) $!" &2)%284%07 &233"!"8%<`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

!"#$ P< G-H'@$)"& )D*!4" 5$/"3D!() :6!D*# $/"!$6" D3 )"/"8 )*@d"4%)B 3D! %1"3D*! 8D2)" &*!$%2D8) :+F J;;()> WF A;;;()> aF =;;;()> eF L;;;()B< ?1")%2(*0*) D8)"% 2) $% ;()< ?1" %!$8)2%2D8 3!D( 8D2)" %D HCT 2) ($!["& @7 $ &D%%"&028"<

! =;;L I"&"!$%2D8 D3 X*!D#"$8 ,"*!D)42"84" TD42"%2")F 1&'($-"2 3(&'2". (4 5-&'(),0-2,-F !"F P;QPfP;O;

P;QR +< T"2%1"!'G!"2)0"! -* ".6

Seither-Preisler et al., 2004

EEG  version  of  POR  (CPR):  IRN  s1muli  

Iteration steps 8 and 32

IRNo: no pitch

only slow spectro- temporal

fluctuations

Two segments 500 ms 250 ms

Krishnan et al., Neuropsychologia (2012)

Extrac1on  of  brainstem  and  cor1cal  pitch  response  

0 20030 Hz

80 Hz

A

C

B

400 600 800

+

+

-

-

1000Time (ms)

0 200 400 1000Time (ms)

-5

0

5

500 550 600 650 700 750 800 850Time (ms)

600 800

Ampl

itude

(μV)

5 μV

0.4

μV

P1

N1

CPR Components

CORTICAL

BRAINSTEM

FFR

3-25 Hz

75-1500 Hz

CPR-Average: EEG down-sampled to 2048Hz; Analysis epoch: 1200 ms including 100 ms prestimulus; 1000 sweeps

Pitch  Onset  Response  is  sensi1ve  to  change  in  pitch  salience  

JJ

J

J

pIRNo nIRNo IRN8 IRN32

610

620

630

640

650

Late

ncy

(ms)

Stimulus

-6

-5

-4

-3

-2

-1

0

Am

plitu

de (u

V)

***

***

Since there are multiple transient components of the POR We chose to name our response as the Cortical Pitch Response (CPR)

Comparison  between  neural  and  behavioral  data  

FFR vs CPR; CPR vs F0 DL; and FFR vs F0 DL are correlated suggesting that behavioral pitch responses were well predicted by both neural responses. These results suggest smaller F0 DLs are associated with more robust encoding at brainstem and cortical levels

CPR:  Experience-­‐dependent  enhancement  

!2

0

2

500 600 650 750 800 850 900!2

0

2

500 550 600 650 700 750 800 850 900!2

!1

0

1

2

!2

0

2

2 µV

+

-100 100 300 500 700 900Time (ms)

0 100 200 300 400Time (ms)

T2_200

T2_250

T2_150

P1

N1

Pa

Na

Pb

Nb

Pc

Nc

Pa

Na

PbNb

Pc

Nc

Pa

Na

Pb

Nb

Pc

Nc

Pa

Na

Pb

Nb

Pc

Nc

C E

No systematic latency effects

For both Na-Pb & Pb-Nb, C>E

Lateraliza1on  only  at  temporal  electrode  sites  

500 550 600 650 700 750 800 850 900!1

!0.5

0

0.5

1

500 550 600 650 700 750 800 850 900!1

!0.5

0

0.5

1

0

1

500 550 600 650

700

750 800 850 900!1

!0.5

0

0.5

1

!1

0

1

500 550 600 650 700 750 800 850 900!1

!0.5

0

0.5

1

25.0

13.5

2.0

Fre

qu

en

cy

(H

z)

1.0

µV

T2_250

T2_200

T2_150

Chinese English

T7

T8

T8

T8

T7

T7

T8

T8

T8

(µV2/Hz)

0.006

0

Pa

Na

Pa

Na

Pb

Nb

Pc

Nc

Pa

Na

Pa

Pa

Na

Na

Pb

Pb

Nb

Nb

Nb

Pc

Pc

Nc

Nc

Time (ms)

100 300 400100400300

T7 T7

T7

2000 200 0 0 200 600 800 1100400 0 200 600 800 1100400

Pb

Pb

PbPa

Na

Nb

Nb

Pc

Pc

Pc

Nc

Nc

Nc

T7 C T8 C T7 E T8 E

Chinese English

500 550 600 650

700

750 800 850 900!1

!0.5

0

0.5

1

!1

0

1

500 550 600 650 700 750 800 850 900!1

!0.5

0

0.5

1

25.0

13.5

2.0

Freq

uen

cy (

Hz)

1.0

µV

T2_250

Chinese English

F3

F4

F4

F4

F3

F3

F4

F4

F4

(µV2/Hz)

0.006

0

Pa

Na

Pa

Na

Pb

Nb

Pc

Nc

Pa

Na

Pa

Pa

Na

Na

Pb

Pb

Nb

Nb

Nb

Pc

Pc

Nc

Nc

Time (ms)

100 300 400100400300

F3 F3

F3

2000 200 0 0 200 600 800 1100400 0 200 600 800 1100400

Pb

Pb

Pb

Pa

Na

Nb

Nb

Pc

Pc

Pc

Nc

Nc

Nc

F3 C F4 C F3 E F4 E

500 550 600 650 700 750 800 850 900!1

!0.5

0

0.5

1

0

1

500 550 600 650 700 750 800 850 900!1

!0.5

0

0.5

1

500 550 600 650 700 750 800 850!1

!0.5

0

0.5

1

500 550 600 650 700 750 800 850 900!1

!0.5

0

0.5

1

!1

0

1

T2_150

T2_200

Chinese English

F3/F4: No asymmetry for either group

T7/T8: Stimulus-dependent rightward asymmetry only for the Chinese

Stimulus-dependent rightward asymmetry for Chinese only for T2-200 and more robust effect for T2-250.

T2_250

0.0

0.5

1.0

1.5

2.0 T7 C T8 C T7 E T8 E

Response ComponentsNa-Pb Pb-Nb

Peak

to P

eak

Ampl

itude

(V)

0.0

0.5

1.0

1.5

2.0

F3 C F4 C F3 E F4 E

Isola1ng  pitch  related  components  from  s1mulus  offset  component  

On  Hierarchical  processing  for  pitch  (PaIerson  et  al.,  2002)  

Ø Stage  1.Extrac-on  of  -me-­‐interval  informa-on  from  neural    firing  paQern  in  the  AN  and  construc-on  of  interval  histograms  in  the  IC  and  Thalamus    

Ø Stage  2.  Determining  the  specific  value  of  pitch  and  its  salience  from  the  interval  histograms-­‐  probably  in  lateral  HG  (creates  summary  histograms  and  locates  the  first  peak)  

 Ø Stage  3.  Determine  con-nuous  (tonal  language)  and  

discrete  (music)  pitch  changes    and  tracking  them-­‐beyond  PAC  in  STG  or  PP.  this  is  were  asymmetries  emerge!  

Mechanisms: Corticofugal pathways in experience-dependent pitch encoding

Experience which engages cerebral cortex shapes subcortical circuitry via corticofugal (CF) modulation

Krishnan & Gandour Brain and Language (2009) 2

IC, inf. colliculus AC, auditory cortex MGB, medial geniculate body LE, left ear RE, right ear CFL, left corticofugal

Mechanisms: Local brainstem mechanisms underlying experience-dependent encoding of pitch-relevant information

Pitch is organized orthogonal to frequency in the IC neurons

English

Chinese Langner (1992,1997)

Temporal correlation analysis model

Krishnan & Gandour Brain and Language (2009)

Best modulation frequency neurons along the pitch axis extract pitch relevant periodicities

Pitc

h

Long-term experience could sharp tuning, improve coherence of neural-phase-locking, and increase synaptic efficiency

Ravi Krishnan Jack Gandour Jayaganesh

Swaminathan* Gavin Bidelman* Chris Smalt* Sharada

Ananthakrishnan* Jill Wendell Suresh Chandan * Students who have

graduated

Auditory Electrophysiology Lab

Our mantra for studying tones

Bob Dylan Rainy Day Women 1966 Everybody must get sTONED!