effects of hearing loss on learning new...

A. Pittman Copyright © 2016 All rights reserved.

EFFECTS OF HEARING LOSS ON LEARNING NEW

INFORMATIONPART 2

What does it take to learn a new word?

Andrea Pittman, PhDArizona State University


MERGING WORD RECOGNITION WITH WORD LEARNING

(Luce & Pisoni, Ear Hear, 1998)(Gray, Pittman & Weinhold, JSLHR, 2014)(Storkel & Lee, Lang Cog Proc, 2011)(Leach & Samuel, Cog Psych, 2007)

Stimulus Input

Acoustic Phonetic

Pattern Activation

Lexical Processing

(Neighborhood

Activation)

Higher Level

Lexical

Information

Known

Configuration Detection

Unknown

Lear

nin

g Recognition


WORD RECOGNITION AND LEARNING IN ADULTS AND CHILDREN WITH HEARING LOSS

Two-year study funded by the Industry Research Consortium

Overall project goal: Determine the benefits of certain amplification features on auditory tasks that increase in cognitive demand.

Current data analyses: Compare the performance of children and adults who have hearing loss.


Children: 20 NH (8-12 years)

21 HL (8-12 years)

Adults: 15 NH (50-67 years)

17 HL (52-78 years)

PARTICIPANTS

-20

0

20

40

60

80

100

120

HEA

RIN

G L

EV

EL (

dB

)

NHAHIA

-20

0

20

40

60

80

100

120

250 500 1000 2000 4000 8000

HEA

RIN

G L

EV

EL (

dB

)

FREQUENCY (Hz)

NHCHIC


HEARING AID CONDITIONS

Amplification

1. Standard amplification (4 kHz bandwidth)

2. High-frequency amplification (10 kHz bandwidth)

Extended bandwidth

or Non-linear frequency compression

Best performance with either type of amplification was included in the analyses.


BINAURAL AIDED HEARING THRESHOLDSCHILDREN

-20

0

20

40

60

80

100

120

250 500 1000 2000 4000 8000

HEA

RIN

G L

EV

EL (

dB

)

FREQUENCY (Hz)

Aided

Unaided

-20

0

20

40

60

80

100

120

250 500 1000 2000 4000 8000

HEA

RIN

G L

EV

EL (

dB

)

FREQUENCY (Hz)

Aided

Unaided

STANDARD HIGH-FREQUENCY


BINAURAL AIDED HEARING THRESHOLDS ADULTS

-20

0

20

40

60

80

100

120

250 500 1000 2000 4000 8000

HEA

RIN

G L

EV

EL (

dB

)

FREQUENCY (Hz)

Aided

Unaided

-20

0

20

40

60

80

100

120

250 500 1000 2000 4000 8000

HEA

RIN

G L

EV

EL (

dB

)

FREQUENCY (Hz)

Aided

Unaided



BINAURAL AIDED HEARING THRESHOLDS ADULTS VS. CHILDREN

-20

0

20

40

60

80

100

120

250 500 1000 2000 4000 8000

HEA

RIN

G L

EV

EL (

dB

)

FREQUENCY (Hz)

-20

0

20

40

60

80

100

120

250 500 1000 2000 4000 8000

HEA

RIN

G L

EV

EL (

dB

)

FREQUENCY (Hz)



TEST PARAMETERS

Testing

53 dB SPL in quiet

0o azimuth

Data collection

Computer interface

Digital audio recordings

Sessions

1 Unaided

1-2 Aided


WORD RECOGNITION

food

poolboat

naglimb

shoutsubvine

goosedime

Stimulus Input

Acoustic Phonetic

Pattern Activation

Lexical Processing

(Neighborhood

Activation)

Higher Level

Lexical

Information

Known

Luce, & Pisoni, (1998) Recognizing spoken words: the neighborhood activation model, Ear & Hearing, 19(1), 1-36.

Recognition


WORD RECOGNITIONBEST PERFORMANCE

Non-linear frequency compression

Wide-band amplification

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100

Hig

h-F

requency

Am

plif

icat

ion (

% c

orr

ect

)

Standard Amplification (% correct)

HIC

HIA


WORD RECOGNITION RESULTS

30

85

41

76

0

10

20

30

40

50

60

70

80

90

100

Unaided Aided

Perf

orm

ance

(%

Corr

ect

)

Listening Condition

ChildrenAdults

p=.29

p=.051


WORD RECOGNITION RESULTS

Recognition of familiar words was the same in these children and adults with hearing loss.

Are they able to detect and learn new words equally well?


AUDITORY LEXICAL DECISION

baby

dopperssoovie

onlygrandma

torsesbeforetaybe

stoobioutside

Stimulus Input

Acoustic Phonetic

Pattern Activation

Lexical Processing

(Neighborhood

Activation)

Higher Level

Lexical

Information

Known


Unknown


AUDITORY LEXICAL DECISION


AUDITORY LEXICAL DECISION BEST PERFORMANCE

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100

Hig

h-F

requency

Am

plif

icat

ion (

% c

orr

ect

)


HIC

HIA




AUDITORY LEXICAL DECISION RESULTS

30

83

33

65

0

10

20

30

40

50

60

70

80

90

100

Unaided Aided

Perf

orm

ance

(%

Corr

ect

)

Listening Condition

ChildrenAdults

p=.76

p<.01


RESPONSE ANALYSIS

Stimulus

Real“cat”

NotReal“glug”



0

10

20

30

40

50

60

70

80

90

100

Unaided Aided

Perf

orm

ance

(%

Corr

ect

)

Listening Condition

ChildrenAdults

Real

Nonsense



Children are better than adults at identifying words they do and do not know when they occur in isolation.

How do they do when theunknown words are embedded in sentences?


NON-WORD DETECTION

Stimulus Input

Acoustic Phonetic

Pattern Activation

Lexical Processing

(Neighborhood

Activation)

Higher Level

Lexical

Information

Known


Unknown


NON-WORD DETECTION

Children (8-12 years)

19 HI Children 29 NH Children

Pittman & Schuett (2013) Effects of semantic and acoustic context on nonworddetection in children with hearing loss, Ear & Hearing, 34(2) 213-220.


NON-WORD DETECTION

Close all three doors.

0 1 2 30


NON-WORD DETECTION

Cooks make hot foom.

0 1 2 31


NON-WORD DETECTION

Overall performance (% correct)

Error analyses

Under-detect

Over-detect

0

10

20

30

40

50

60

70

80

90

100

PER

CET

NT

OF T

OTA

L

OVER

UNDER

CORRECT

REPAIR

MISPERCETION


NON-WORD DETECTION

0%

20%

40%

60%

80%

100%

NH HL NH HL

OVER

UNDER

CORRECT

QUIET

PER

CEN

T O

F T

OTA

L

STEADY-STATE NOISE

MISPERCEPTION

REPAIR

CORRECT

Pittman & Schuett (2013) Effects of semantic and acoustic context on nonworddetection in children with hearing loss, Ear & Hearing, 34(2) 213-220.


NON-WORD DETECTIONBEST PERFORMANCE

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100

Hig

h-F

requency

Am

plif

icat

ion (

% c

orr

ect

)


HIC

HIA




NON-WORD DETECTIONRESULTS

41

68

35

63

0

10

20

30

40

50

60

70

80

90

100

Unaided Aided

Perf

orm

ance

(%

Corr

ect

)

Listening Condition

ChildrenAdults

p=.36

p=.27


NON-WORD DETECTIONRESULTS

41

68

35

63

0

10

20

30

40

50

60

70

80

90

100

Unaided Aided

Perf

orm

ance

(%

Corr

ect

)

Listening Condition

ChildrenAdults

16

20

22

10

REPAIR

MISPERCETION


CONCLUSIONS

Children and adults are equally good at identifying words they do and do not know when they occur in context. However, children tend to repair nonsense words while adults tend to misperceive real words.

Why?


COMPETING STRATEGIES

Crukley et al (2011) An exploration of non-quiet listening at school, J Ed Aud 17, 23-35


NEXT QUESTION

If detection of new words is not an issue, how well can adults learn new words compared to children?


RAPID WORD LEARNING

Stimulus Input

Acoustic Phonetic

Pattern Activation

Lexical Processing

(Neighborhood

Activation)

Higher Level

Lexical

Information

Known

Configuration

Unknown

Lear

nin

g


RAPID WORD LEARNING


RAPID WORD LEARNING

0

10

20

30

40

50

60

70

80

90

100

5 25 45 65 85 105 125 145

PER

FO

RM

AN

CE (

%)

TRIALS

𝑃𝑐 = 1 − 0.8𝑒−𝑛/𝑐


0

10

20

30

40

50

60

70

80

90

100

5 25 45 65 85 105 125 145

PER

FO

RM

AN

CE (

%)

TRIALS

Average

Arithmetic Mean

RAPID WORD LEARNING

Averaged Data

Averaged Fits


WORD LEARNING AND BANDWIDTH

AGE (years)

HEA

RIN

G L

OSS

(d

egr

ee)

Children26 HI Children 41 NH Children

Pittman (2008) Short-term word-learning rate in children with normal hearing and children with hearing loss in limited and extended high-frequency bandwidths JSLHR, 51(3) 785-797.


WORD LEARNING AND BANDWIDTH

0

10

20

30

40

50

60

70

80

90

100

5

15

25

35

45

55

65

75

85

95

105

115

125

135

145

PE

RF

OR

MA

NC

E (

%)

TRIALS

NHC (9k Hz)

HIC (9k Hz)

HIC (4k Hz)

Pittman (2008) Short-term word-learning rate in children with normal hearing and children with hearing loss in limited and extended high-frequency bandwidths JSLHR, 51(3) 785-797.


WORD LEARNINGBEST PERFORMANCE

0

1

2

3

0 1 2 3

Hig

h-F

req

Am

plif

icat

ion (

lear

nin

g sp

eed)

Standard Amplification (learning speed)

HIC

HIA




RAPID WORD LEARNING

0

20

40

60

80

100

5 15 25 35 45 55 65 75 85 95 105115

Perf

orm

ance

(%

Corr

ect

)

Trials

Unaided

Aided

p=.78

p=.31

LEARNING SPEED FOR WORDS ONLY (no PLURALS)


RAPID WORD LEARNING

0

20

40

60

80

100

5 15 25 35 45 55 65 75 85 95 105115

Perf

orm

ance

(%

Corr

ect

)

Trials

No Plurals

All Words

p<.01

LEARNING SPEED FOR WORDS ONLY (with PLURALS)


RAPID WORD LEARNING

Children and adults are equally good at learning new words when provided with optimal amplification.


NEXT QUESTION

Are words learned rapidly retained?


PERCEPTUAL LEARNING OF WORDS

Purpose:

Examine the relationship between performance for the word learning task on one day and the participant’s retention of those words the next day.

Manipulated the number of training trials.

Pittman, Wright, Wright & Latto (in progress)


PERCEPTUAL LEARNING OF WORDS

Participants:

98 normal-hearing undergraduates (130+ total)

Method:

Day 1: Learned 3 sets of 5 nonsense words

Day 2: Completed an online post-test



WORD LEARNING PARADIGM



POST TEST

Administered online

Enter (type) the name learned for each image.

The images are randomly distributed.

The word bank contains the 15 words earned as well as 15 orthographic/phonetic foils.



RELATION BETWEEN LEARNING AND RETENTION

0.0

0.2

0.4

0.6

0.8

1.0

20 30 40 50 60 70 80 90 100 110 120 130 140

Perf

orm

ance

(p

rop

ort

ion

co

rrec

t)

# of Training Trials

Training

Retention

Time on Task (minutes)

2 4 6 8



CONCLUSIONS

Word learning is not a perishable skill, adults can learn as quickly as children.

Hearing loss impairs/slows word learning but normal or near-normal word learning can be achieved with optimal amplification.

Words learned rapidly are retained for at least a day.


ISN’T SPEECH PERCEPTION ENOUGH?

0

1

2

3

0 0.2 0.4 0.6 0.8 1

Lear

nin

g Sp

eed

Word Recognition

HIA

HIC


smegs pum fleep gorn soovie nunny grazdamorses ranfed smiving hwiking brettyekleep wattem gorning kello doppersstoobie taybe wrugger rin pomen tunclemooking zigger winber dwendy samapasker franding sishing wattem tanti unelmanka yosie batrol kissle rathit dravecloums loctor the end higger cougin beelstrouzle mayfe kerkle hatty riger thocketprozen tieces transa finter mivshen katherpeagie henkred vummy rumpdin pimi ifstidranfed tryving brenny ricken potton ofliropim voing bipping samder mither sasepaltifer smiving chamer beefar etote chorketlelling veking geaunen piving poctor kinnerkidnesh sefond sama dazi ketting rouplevello tryving yosie rin bretty ranfed smegspasker

effects of hearing loss on learning new...

Documents