Nessun titolo diapositiva

1
How spatial attention modulates reading aloud and
lexical decision Evidence from Italian neglect
dyslexia patients
Lisa S. Arduino University of Milano-Bicocca,
Milan ISTC-CNR and Fondazione S. Lucia, IRCCS,
Rome Cristina Burani Institute of Science and
Technology of Cognition ISTC-CNR, Rome Giuseppe
Vallar University of Milano-Bicocca
The Third International Conference on the Mental
Lexicon Banff, Alberta, Canada October 6-8,
2002.
2
NEGLECT DYSLEXIA (ND)
UNILATERAL SPATIAL NEGLECT disturbance in
perceiving, representing and orienting attention
to the controlesional side of space.

• LESION RIGHT INFERIOR PARIETAL LOBULE
  (Bisiach Vallar, 2000 Vallar et al.,
  1998)
• NEGLECT DYSLEXIA SINGLE WORD READING
  (egocentric coordinate frames)
• TARGET ALBERO
  tree (Ellis et al. 1987)
• SUBSTITUTION POBERO
• OMISSION BERO
• ADDITION COSBERO

3
DISSOCIATIONS

• Làdavas et al. (1997, Neuropsychologia) Simple
  words and nonwords presented centrally (9
  patients)
• POOR READING ALOUD
• BUT
• PRESERVED LEXICAL DECISION AND
• SEMANTIC JUDGEMENT

• Vallar et al. (1996, Journal of Clinical and
  Experimental Neuropsychology)
  compound words (E.S.)

camposanto cemetery campo field santo saint
x
camposanto
Severe ND in reading aloud BUT appropriate
association (e.g., coffin)
4
Explanations

• Reading aloud differs from lexical decision
  (semantic judg. and associations) for
• Diffculty lexical decision is easier than
  reading aloud and requests less information from
  the left side (guessing strategy).
• The different involvement of spatial co-ordinate
  frames (Vallar et al., 1996).
• The differential use of reading routes (Ladavas
  et al., 1997) DRC model (Coltheart et al., 2001).

5
Written stimulus
The route operates on the whole word-
form NO attentional scanning
The route operates serially attentional
scanning from left-to-right
ORTHOG. LEXICON
G P C rules
Semantics
PHONOL. LEXICON
Phonemic buffer
OUTPUT
6
The present study

• Aimed at specifying in further detail the
  preserved lexical processing in patients with
  left ND by exploring in LD tasks, the effect of
  morpho-lexical variables, which influence the
  performance of Italian unimpaired subjects.

7
EXPERIMENT 1Morphologically simple words and
nonwords

• Dissociations between reading aloud (RA) and LD
  in neglect dyslexia patients the same stimuli
  presented to six patients for both RA and LD
  (Arduino et al., 2002, Cognitive
  Neuropsychology). Untimed presentation.
• LD accuracy The six patients were compared to 12
  controls (matched for age, sex and educational
  level)
• Lexical effects in LD four patients LD
  performance was compared to that of non
  neurological younger adults. Timed presentation
  (500 or 700 ms.)

8
EXPERIMENT 1
LIST 240 simple words and nonwords
DEPENDENT VARIABLE errors
PROCEDURE untimed (all) timed (4 patients)
A) 40 WORDS High and Low surface frequency
(50). B) 72 BISYLLABIC NONWORDS (5-6 letters).
Neighborhood frequency (High/Low) BRISI
CRISI NERPE SERPE
9
Experiment 1 RA and LD patients
errors
Experiment 1 LD patients vs. controls
errors
10
ESP. 1LD with timed presentation (500 ms.) 4
patientsHigh and Low frequency words correct
answers.

• High-frequency words are recognized faster and
  with less errors than low-frequency words
  (Colombo, 1992, JEPHPP Burani et al., 2002,
  Brain and Language)

11
EXP. 1 LD with timed presentation (500 ms.) 4
patientsNonwords with High/Low frequency
neighbor errors
BRISI CRISI NERPE SERPE
12
LD non neurological subjects (Arduino Burani,
accepted, JPR)

• Stimuli the same
• Participants 49 university students
• Dependent variable RT and errors

Error analysis showed the same pattern
13
EXPERIMENT 2Morphologically complex words and
nonwords

• Dissociation between RA and LD in neglect
  dyslexia patients the same stimuli presented to
  six patients for both RA and LD (Arduino et al.,
  2002). Untimed presentation.
• LD accuracy The six patients were compared to 12
  controls (matched for age, sex and educational
  level)
• Lexical effects in LD three patients LD
  performance compared to non neurological younger
  adults. Timed presentation (700 ms.)

14
EXP. 2
LIST 300 morphologically complex words and
nonwords
DEPENDENT VARIABLE errors
PROCEDURE untimed (all) timed (3 patients)

• A) 88 suffixed derived words (Burani Thornton,
  2002, Linguistics). All words were low frequency
• 44 with HF root (CONSUM-ISMO consumerism)
• 44 with LF root (SIMBOL-ISMO simbolism)

• B) 138 nonwords (Burani et al., 1997, Yearbook
  of Morphology Burani et al., 1999, Brain and
  Language)
• LAMPAD-ISTA (RS)
• RONDIN-OSTO (RS-)
• ROVOLL-ISMO (R-S)
• MEVIN-OSTO (R-S-)

15
Exp. 2 RA and LD patients errors
Exp. 2 LD patients vs. controls errors
16
EXP. 2LD with timed presentation (700 ms.)
suffixed derived words

• Burani Thornton (2002) less errors in deciding
  upon words with high-frequency root.

17
EXP 2LD with timed presentation (700 ms.)
morphologically complex nonwords

• Burani et al. (1997, 1999) Burani Thornton
  (2002)
• more errors on nonwords that included either one
  or two constituent morphemes with respect to
  nonwords with no morphemes

18

• Summary

• The results of both experiments confirmed that
  neglect dyslexia patients lexical decision
• is preserved compared to reading aloud
• is normal compared to the performance of control
  subjects
• Moreover the results show that LD
• is affected by the same morpho-lexical
  characteristics that influence non neurological
  younger adults
• is not related to the severity of neglect
  dyslexia

19
Exp. 1
Exp. 2
20
CONCLUSIONS

• Guessing strategy The fact that morpho-lexical
  effects also emerged in the patients LD allows
  us to discard the hypothesis that the patients
  adopt a rough guessing strategy in LD.
• Differential use of the reading routes (Ladavas
  et al., 1997)
• LD good performance because patients made use
  of the lexical route (no serial processing is
  required)
• RA impaired performance because patients made
  use of the sublexical route (serial processing,
  from left-to-right)
• Moreover
• For some Italian patients the lexical route is
  available for reading aloud (Arduino et al.,
  2002). It is the availability of the lexical
  route, which makes use of the whole word-form,
  that allows the patients to process the stimuli
  correctly.

21

• Arduino et al.s data (2002) may be taken as
  further evidence that when patients may have
  access to the entire word-form directly, through
  the lexical route, their disturbance is
  ammeliorate because this latter procedure does
  not require a sequential, from left-to-right,
  processing.
• In conclusion
• The dissociation between reading aloud and
  lexical decision may be due to the fact that
  reading aloud requires, at different processing
  stage, a left-to-right sequential processing that
  is impaired in neglect patients, whereas it is
  not required in LD.

22

• Some authors have suggested that word processing
  may involve two anatomically distinct attentional
  structures
• A posterior attentional system which is
  devoted to the allocation of visual spatial
  attention across the visual field (necessary for
  reading aloud, and which is impaired in neglect
  patients) and a more central anterior attentional
  system (preserved in neglect patients) which
  plays a role in lexical/semantic access (see
  Carr, 1992, American Journal of Psychology, for a
  review).

23
READING ALOUD (ARDUINO ET AL, 2002)
24
Written stimulus
The route operates on the whole word-
form NO attentional scanning
The route operates serially attentional
scanning from left-to-right
ORTHOG. LEXICON
G P C rules
Semantics
PHONOL. LEXICON
Phonemic buffer
OUTPUT
25
(No Transcript)
26
(No Transcript)
27
Reading test (Vallar et al., 1996)
neglect errors out of the total number of errors
28
Exp. 1 HIGH\LOW FREQUENCY WORDS
Exp. 1 NONWORDS WITH HIGH\LOW FREQUENCY NEIGHBOR
29
Exp. 1Percent of neglect errors as a function
of error type
30
EXPERIMENT 2Reading aloud morphologically
complex words and nonwords
31
Exp. 2 Percent of neglect errors in reading
word and nonword targets
32
Exp. 2 DERIVED (SUFFIXED) WORDS HH
BASS-EZZA LL BEFF-ARDO
Exp. 2 MORPH. COMPLEX NONWORDS RS
LAMPAD-ISTA R-S- MEVIN-OSTO
33
Exp. 2Percent of neglect errors as a function
of error type
34
RESULTS

• FIVE PATIENTS SHOWED LEXICAL EFFECTS IN READING,
  WHILE ONE PATIENT DID NOT (A.A.)
• FEW ERRORS IN READING
• words vs. nonwords (Exp. 1 and 2)
• high vs. low-frequency words (Exp. 1)
• nonwords with no high-frequency neighbor (Exp.1)
• derived words with high-frequency constituents
  (root and suffix). (Exp. 2)
• morph. complex nonwords with real root and suffix
  (Exp. 2)

35
CONCLUSIONS
The two types of neglect dyslexia are
different manifestations of a single attentional
disorder, different in degree.
Relationship between the severity of the
attentional disturbance and the presence vs.
absence of lexical effects in reading.
BUT Relationship between the severity of left
neglect and lexical effects is specific to the
domain of neglect dyslexia, and not extending to
other manifestations of the disorder.
36
Percentage of reading errors committed by the two
patients under condition of unconstrained time
(Exp. 1 and 2)
37
These results suggest that neglect dyslexia
reflects a form of impairment in the spatial
allocation of attention or in spatial
representation, specific to the domain of the
reading system. By and in line with this view,
neglect dyslexia has been described in the
absence of other manifestations of neglect
symptoms for nonverbal material (Bisiach et al.,
1990), or involving the one side of space
opposite to the one where neglect for nonverbal
material is present (Cubelli et al., 1991
Riddoch et al., 1995).
38
(No Transcript)
39
EXP. 1 (6.0)
Mean stimuli length
EXP. 2 (8.3)
40

• Within a left-to-right gradient interpretation
  of left neglect
• the assumption can be made that the longer is the
  letter string
• the more degraded is the internal representation
  of its left side.
• The increase in omissions with longer letter
  strings represents
• a counterpart, in the reading domain of the well
  known effect
• of line length in segment bisection The
  rightward shift of the
• subjective midpoint increases with longer lines
  (Vallar et al.,
• 2000 Bisiach et al., 1983).
• Within this interpretative framework the more
  material is to
• be computed on the left side of the letter
  string, the greater is
• the probability of a defective processing, that
  is of an omission
• error.

41
Length effect. Percentage of neglect errors to
5-6 vs. 7-11 letter targets (data from Exp. 1 and
2).
42

• The relationship between error types (sub. vs.
  omiss.)
• the severity of the attentional disorder and
  lexical
• effects also falls along a continuum
• Large majority of omissions may be associated
  with
• a more severe attentional disorder and with the
• absence of lexical effects in reading.
• A large majority of substitutions may be
  associated
• with preserved lexical effects and a less severe
• attentional disorder.