Semantic and gender priming in Frontotemporal Dementia

Semantic and Gender Priming in Frontotemporal Dementia

Claudia REPETTOa,1,, Rosa MANENTIb, Stefano CAPPAc , Carlo MINIUSSIb,d and
Giuseppe RIVAa,e
aIstituto Auxologico Italiano, Applied Technology for Neuro-Psychology-ATNP
Lab, Milan, Italy
bIRCCS S. Giovanni di Dio-Fatebenefratelli, Brescia, Italy
cVita-Salute San Raffaele University, Milan, Italy
dDepartment of Biomedical Sciences and Biotechnologies, University of
Brescia, Brescia, Italy
,eUniversità Cattolica del Sacro Cuore, Interactive Communication and
Ergonomics of New Technology – ICE-NET Lab., Milan, Italy

Abstract. Modifications of language processing can be observed both
in normal ageing and in the most common forms of degenerative
dementia, such as Alzheimer's disease and the spectrum of
frontotemporal dementias. The present experiment tests at the same
time semantic and syntactic aspects of language processing in
patients with frontotemporal dementia, using an on-line paradigm that
allows researchers to evaluate the real linguistic competence of the
patients.

Keywords. Semantics, gender, priming, frontotemporal dementia

Introduction

On-line processing tasks, such as priming paradigms, are ideally suited to
assess modifications of cognitive processing associated with normal ageing
and early dementia. In fact, priming is largely based on automatic
processes, that enable to bypass execution deficits, such as slowed motor
response, proper to healthy older people and patients suffering from
cognitive deterioration. Semantic priming is the facilitation induced by a
prior activation of items semantically related to the target, and results
in faster responses to the target if compared to a baseline condition
(without any semantic cue). In gender priming the facilitation is due to
the gender agreement between the prime and the target. Gender priming is
considered a particular type of syntactic priming. Semantic and syntactic
priming have been studied on young people by Bentrovato [1]: they found a
large, additive?? facilitation effect of both gender and semantics. These
results were taken to indicate that normal subjects could take advantage of
both source of information, with no significant interference when they are
both discordant. The same procedure, than, has been applied in a sample of
patients with diagnosis of Alzheimer's Disease (AD), compared with a sample
of normal ageing subjects [2]. Researchers found effects of facilitation
due to the concomitant congruence of both gender and semantics in both the
groups; by opposite, a striking difference between controls and patients
was discovered, when both information (semantic and gender prime) were
discordant. In this case an interference effect appeared, suggesting that
AD patients could feel the effect of difficulties in inhibiting irrelevant
information. The present study aimed to test gender and semantic priming in
a sample of patients suffering from different variants of frontotemporal
dementia, a degenerative disease in which language processes could be
affected at different stages.
1. Methods
Twenty-four patients with diagnosis of frontotemporal dementia, who
received a complete neurological and neuropsychological evaluation, entered
this study. The test used was the Italian Word Reading Test [1]: it is
featured by brief two-sentences stimuli, orally presented, with a visually
presented target word. Auditory sentences provide the semantic and gender
context and represent the primes. The sentence context ends with an
article, that provides gender cue and precedes the target word. Different
priming conditions have been created manipulating semantic and gender
agreement: semantic concordance/discordance (S+/ S-) between the auditory
context and the target word establishes the semantic priming relationship;
gender agreement/disagreement (G+/ G-) between the article and the target
word gives rise to gender priming conditions. Experimental protocol, thus,
comprised a 2 (S+/ S-) X2 (G+/ G-) design, and an additional neutral
condition, that served as baseline, without any semantic or gender cue
(NN). Table 1 illustrates examples of the sentences used in each condition.
Subjects were seated in front of a computer screen, wearing headphones
with a microphone, that were connected to the sound amplifier port of the
computer. Response times were collected in milliseconds through the
Carnegie Mellon Button Box, a measuring device including a voice key with
an external time crystal and 1 ms resolution, designed for use with a
Macintosh Computer. Experimental design was implemented on Psyscope [3].
Subjects were instructed to listen carefully the auditory sentence
context looking at the fixation point (++) on the screen and, when the
sentence was halted, to read as quickly as possible the word appeared in
the place of the fixation point.
Table 1. Examples of prime-target pairs for each condition
"Prime " "Target "Prime "Condition "
"Quando vado a " " "per questo mia mamma " "
"letto prima di " " "mi ha regalato una " "
"addormentarmi "UN (masc) "LIBRO "collezione di romanzi "G+S+ "
"leggo sempre "UN (masc) "(book) "gialli "G+S- "
"(when I go to "UNA (fem) "TOPO " "G-S+ "
"bed before "UNA (fem) "(mouse) " "G-S- "
"falling asleep " "LIBRO " " "
"I always read) " "(book) " " "
" " "TOPO " " "
" " "(mouse) " " "
"Silvia ha fatto" "LIBRO "per 5 volte. "NN "
"un test in " "(book) "L'insegnante ha detto " "
"inglese: doveva" " "che è l'unico modo per" "
"ripetere " " "migliorare la " "
"(Silvia took a " " "pronuncia (five times." "
"test in " " "The teacher said that " "
"English: she " " "was the only way to " "
"had to repeat) " " "improve her " "
" " " "pronunciation) " "
" " " " " "


2. Results


In order to assess the effect of gender and semantic congruence a 2X2 Anova
within-subjects was conducted. The results indicated an effect of gender
[F(1,23) = 5,608; p = 0,023] and semantics [F(1,23) = 14,567; p = 0,001],
while the interaction gender X semantics was not significant [F(1,23) =
3,810; p = 0,063]. At this point it was important to determine the
direction of priming; to do that, we compared reaction times of each
condition with the ones of the neutral baseline. Four independent samples t-
test were conducted: a significant facilitation was observed for the G+S+
condition [t (23)= 3,706; p= 0,001]; no other condition reached
significance, even if G-S+ and G-S- showed a trend close to significance in
opposite effect direction (respectively facilitation and interference) [G-
S+: t (23)= 1,913; p=0,068], [G-S-: t (23)= -1,900; p=0,069]. See graph 1to
see facilitation and interference effects of both sample (FTD and AD),
comparing the present results to the ones obtained by Manenti et al (2004)
[2].













Graph 1. The bars indicate the difference of each condition to the neutral
baseline, within each group, thus indicating the presence of facilitation
or interference


3. Conclusion


The present findings indicate that frontotemporal patients behave similarly
to AD patients for what concern language processing: in fact in both
groups, a significant facilitation due to concomitant congruence of gender
and semantics was found. The lack of interference effect in FTD when there
is a complete incongruence (G-S-) is in contradiction with one of the main
feature of frontal lobe disease, that is the difficulty to inhibit
irrelevant information; this result could be explained considering the
small size of sample.



References
[1] S. Bentrovato, A. Devescovi, S. D'Amico, N. Wicha and E. Bates, The
effect of grammatical gender and semantic context on lexical access in
Italian using a timed word-naming paradigm, J Psycholinguist Res 32 (2003),
417-30.
[2] R. Manenti, C. Repetto, S. Bentrovato, A. Marcone, E. Bates and S. F.
Cappa, The effects of ageing and Alzheimer's disease on semantic and gender
priming, Brain 127 (2004), 2299-306.
[3] J. Cohen, B. MacWhinney, M. Flatt and J. Provost, Psyscope: an
interactive graphic system for designing and controlling experiments in the
psychology laboratory using Macintosh computers, Behaviour Research
Methods, Instruments and Computers 25 (1993), 257-271.