PET Count

1
Lexical and Sublexical Components of Age-Related
Changes in Neural Activation During Visual Word
Identification
Whiting, W.L., Madden, D.J., Langley, L.K.,
Denny, L.L., Turkington, T.G., Provenzale, J.M.,
Hawk, T.C., Coleman, R.E. Center for the Study
of Aging and Human Development, and Departments
of Psychiatry and Radiology Duke University
Medical Center, Durham, NC 27710
Word Frequency Activation Effects
Introduction

• Word frequency had a larger effect on RT than
  length for both younger (? -.497,R2 .168 vs.
  ? .101, R2 .007) and older adults (?
  -.457, R2 .146 vs. ? .241, R2 .041) .
• To determine whether the pattern of frequency and
  length effects was statistically different for
  younger and older adults, we used frequency and
  length values to predict RTs for each
  participant. Each participants frequency and
  length coefficient was submitted to an Age x
  Predictor ANOVA.

Striate Cortex
Inferior Temporal Gyrus
Previous studies of word identification have
found that the contribution of lexical/semantic
and sublexical/nonsemantic processing to word
identification is different for younger and older
adults (e.g., Allen et al., 1991 Spieler
Balota, 2000). Spieler and Balota, like others,
found that the more frequently a word appears in
printed text, the more rapidly both younger and
older adults were able to name words. However,
this lexical variable (word frequency) accounted
for a greater proportion of variance in older
adults naming latencies relative to younger
adults. This finding along with other findings
of equivalent word frequency effects for younger
and older adults during word identification
provides evidence of preserved semantic
processing with age. There is also evidence to
suggest that the relationship between sublexical
word properties (e.g., word length) and word
identification RT also varies with age (Allen et
al.). The goal of the present study was to
determine whether age differences in the effects
of lexical and sublexical processes on word
identification might reflect different patterns
of neural activation. Madden et al. (in press)
previously examined the relationship between word
identification RT and activation, but found no
age differences in the correlation between these
two variables. The present study reexamined the
Madden et al. data to see if age differences in
the relationship between RT and activation might
be mediated by lexical and sublexical word
properties.
Word Frequency Coefficient
Younger Adults
1
0.30
Word Frequency Coefficient
Older Adults
Younger
-.263
-.249
0.20
Older
Regression Coefficient
PET Count
PET Count

• Word Frequency effects (coefficients) were
  reliably related to activation in both the
  striate and ITG for older adults only.
• For older adults, greater activation in both the
  striate and ITG was associated with larger
  effects of frequency on lexical decision RT
  (i.e., larger frequency coefficients).

0.10
.133
.050
0
Word Frequency
Word Length
Word Length Activation Effects

• As shown above in Figure 1, although word
  frequency coefficients were similar for older and
  younger adults, word length had a larger effect
  on older adults lexical decision RTs relative to
  younger adults.

Striate Cortex
Inferior Temporal Gyrus
Method

• Participants
• 12 younger adults (M 23.6 yrs, range 20-29 yrs)
  and 12 older adults (M 65.0 yrs, range 62-70
  yrs).
• Design and Procedure
• Word Identification task
• A standard yes/no lexical decision task during
  each trial block, participants were randomly
  presented with either words (e.g., FOREST) or
  pronounceable nonwords (e.g., TWEAL).
• On a single trial, participants pressed one of
  two buttons on a response box to indicate whether
  the letter string was a word or nonword.
• Participants performed 9 blocks of trials
  containing a total of 720 words and nonwords.
• Words Ranged in Frequency from 20 to 65 SFI and 4
  to 10 letters in length within each block.
• Baseline task
• For the baseline/subtraction task, participants
  searched letter strings of Ts and Zs in
  search of a lower case c (e.g., TZTZTZcT) and
  pressed one of two buttons to indicate the
  presence or absence of a c.
• Positron Emission Tomography (PET)
• Measurement of regional cerebral blood flow
  (rCBF) was conducted with a GE Advance whole-body
  PET scanner (35 imaging planes separated by 4.25
  mm). Intrinsic in-plane and axial spatial
  resolutions were approximately 5 mm.
• Radiotracer administration was intravenous bolus
  injection of 10 mCi of H215O. Data acquisition
  was performed in the 3D mode (septa out). Each
  task condition was performed during a separate
  PET scan. There was a total of 12 emission scans,
  three scans for each of four task conditions (1
  baseline, 3 lexical decision).

Younger Adults
Activation Results
Word Length Coefficient

• We next examined whether activation in Madden et
  al. might be related to these lexical and
  sublexical variables.
• Madden et al. found that, relative to the
  baseline search task, word identification was
  associated with greater activation of the
  inferior temporal gyrus (ITG) for older adults,
  but greater visual striate activation for younger
  adults (see Figure 2 below). Our analyses focused
  on these 2 areas of activation.
• We simultaneously regressed activation (PET
  counts) in these 2 areas onto word frequency and
  length coefficients in separate analyses for each
  lexical/sublexical variable and age group (see
  Figures 3 4).

Older Adults
Word Length Coefficient
PET Count
PET Count

• Word Length effects (coefficients) were reliably
  related to activation only in the striate cortex
  of older adults.
• For older adults, lower activation in the striate
  was associated with larger effects of length on
  lexical decision RT.

Older Adults
2
Younger Adults
Conclusions

• For older adults, greater activation in the
  striate and ITG was associated with less
  efficient lexical (word frequency) processing, as
  those with larger coefficients had slower RTs.
• The opposite effect was found for word length
  lower striate activation was associated with less
  efficient sublexical/perceptual (word length)
  processing.
• This pattern of effects may indicate that older
  adults compensate for early visual processing
  deficiencies during word identification by
  engaging in more semantic processing hosted by
  the inferior temporal gyrus.

Behavioral Results

• In the first analysis, lexical decision RTs were
  averaged across participants to obtain mean RTs
  for each of the 720 stimulus words. Word
  frequency and length were then regressed onto RT.
• For both age groups, word frequency coefficients
  were negative indicating that more familiar words
  were identified more quickly. Word length
  coefficients were positive reflecting slower
  responses to longer length words.

References Allen, P. A., Madden, D. J.,
Crozier, L. C. (1991). Adult age differences in
letter-level and word-level processing.
Psychology and Aging, 6, 261-271. Madden, D. J.,
Langley, L. K., Denny, L. L., Turkington, T. G.,
Provenzale, J. M., Hawk, T. C., et al. (in
press). Adult age differences in visual word
identification Functional neuroanatomy by
positron emission tomography. Brain Cognition.
Spieler, D. H., Balota, D. A. (2000). Factors
influencing word naming in younger and older
adults. Psychology and Aging, 15,
225-231. Acknowledgements This work was supported
by National Institute on Aging grants T32
AG00029-26 and R01 AG11622.