Anterior temporal lobes mediate semantic
representation: Mimicking semantic dementia
by using rTMS in normal participants
Gorana Pobric, Elizabeth Jefferies, and Matthew A. Lambon Ralph*
Neuroscience and Aphasia Research Unit, School of Psychological Sciences, University of Manchester, Manchester M13 9PL, United Kingdom
Edited by Edward E. Smith, Columbia University, New York, NY, and approved October 15, 2007 (received for review August 6, 2007)
Studies of semantic dementia and PET neuroimaging investiga-
tions suggest that the anterior temporal lobes (ATL) are a critical
substrate for semantic representation. In stark contrast, classical
neurological models of comprehension do not include ATL, and
ATL, reinforcing the classical view. Using a novel application of
over the ATL, we demonstrate that the behavioral pattern of
semantic dementia can be mirrored in neurologically intact partic-
ipants: Specifically, we show that temporary disruption to neural
processing in the ATL produces a selective semantic impairment
leading to significant slowing in both picture naming and word
comprehension but not to other equally demanding, nonsemantic
repetitive transcranial magnetic stimulation ? semantic cognition ?
jects, and faces. In addition to underpinning comprehension, it
also allows us to express knowledge in a wide variety of domains,
both verbal (e.g., naming and verbal definitions) and nonverbal
(e.g., drawing and object use). As such, it is integral to our
everyday lives, and impairments of semantic memory are ex-
tremely debilitating. Key questions for neuroscience research,
and how do they function?
Various neurological disorders cause impairments of semantic
processing; however, the purest syndrome is semantic dementia
(SD; the temporal lobe variant of frontotemporal dementia) (1).
This neurodegenerative disease results in relatively focal atrophy
and hypometabolism of the anterior temporal lobes (ATL)
bilaterally (2, 3). SD is characterized by progressive impairment
of verbal and nonverbal semantic tasks, with anomia as the first
presenting symptom (4–6). Strikingly, other aspects of language
and cognition remain largely intact. SD patients have increasing
difficulty distinguishing concepts from their semantic neighbors,
reflecting an increasing loss of ‘‘semantic acuity’’ (6, 7). As such,
the patients have greater difficulty activating specific semantic
information (e.g., ‘‘zebras have stripes’’) than more general
properties (e.g., ‘‘zebras are animals’’) (6, 8). Likewise, their
naming difficulties are graded by specificity (higher naming
accuracy for basic-level concepts such as dog than specific ones,
e.g., springer spaniel) (9) with errors reflecting more general
semantic knowledge (e.g., dog 3 ‘‘animal’’).
Careful and extensive assessment of SD patients indicates that
bilateral anterior temporal lobe regions support the formation of
amodal semantic representations. Accordingly, SD patients ex-
hibit poor comprehension of items presented in every modality,
including spoken and written words, pictures, environmental
sounds, smells, and touch (4, 10, 11). The marked semantic
deficit is also apparent in production tasks, such as picture
naming (5), verbal definitions (12), object drawing (13), and
object use (14). The singular, amodal nature of the anterior
emantic memory encompasses the meaning of all types of
verbal and nonverbal stimuli including words, pictures, ob-
temporal lobe system is underscored by the fact that SD patients
show very high correlations between their scores on different
semantic tasks and strong item-specific consistency across mo-
dalities (6, 15).
The anterior temporal lobes are ideal for forming amodal
semantic representations because they have extensive connec-
tions with cortical areas that represent modality-specific infor-
mation (16) (see also the theory of ‘‘convergence zones’’ in ref.
17). Accordingly, Rogers et al. (6) implemented a computational
model of this anterior temporal lobe system in which semantic
representations were formed through the distillation of infor-
mation required for mappings between different verbal and
nonverbal modalities. When damaged, the model reproduced
the behavioral performance of SD patients across a wide variety
of semantically demanding receptive and expressive tasks.
Although the data arising from semantic dementia clearly
implicate the temporal poles, bilaterally, in semantic represen-
research on semantic memory (18–20). Several factors probably
account for this situation. First, classical aphasiological models
have never associated anterior temporal lobe regions with
comprehension disorders—patients with Wernicke’s aphasia
typically have damage to the left posterior middle temporal and
superior temporal gyri, whereas patients with transcortical sen-
sory aphasia have damage to the left temporoparietal or pre-
frontal cortices (21). Second, functional MRI (fMRI) studies of
semantic tasks rarely activate anterior temporal lobe regions but,
in line with the aphasiological models, find activation in left
temporoparietal and prefrontal regions (22, 23). Third, after
unilateral resection of the temporal pole, epilepsy patients do
not have semantic impairment, or at least not to the same degree
as SD patients (24).
Recent studies indicate, however, that these observations are
not contradictory with the results from semantic dementia. First,
direct comparisons of SD and aphasia-related comprehension
impairments show that, although both conditions can lead to
impairment of multimodal semantic cognition (i.e., impaired
semantically driven behavior across verbal and nonverbal mo-
dalities), there is a qualitative difference between the patient
groups; SD results from a gradual dissolution or ‘‘dimming’’ of
the semantic representations themselves whereas aphasic pa-
tients with multimodal comprehension disorders have impair-
ment to the mechanisms that control or shape the activation of
task-relevant information rather than damage to semantic
knowledge per se (7). This indicates that semantically driven
Author contributions: G.P., E.J., and M.A.L.R. designed research; G.P., E.J., and M.A.L.R.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
*To whom correspondence should be addressed. E-mail: matt.lambon-ralph@manchester.
© 2007 by The National Academy of Sciences of the USA
www.pnas.org?cgi?doi?10.1073?pnas.0707383104 PNAS ?
December 11, 2007 ?
vol. 104 ?
no. 50 ?
behavior (which we have come to refer to as ‘‘semantic cogni-
tion’’) comprises two key, interacting components: (i) the core
amodal semantic representations and (ii) ‘‘semantic control,’’
executive-control mechanisms that interact with the underlying
semantic representations to produce task- and time-appropriate
activation of key knowledge for the specific task in hand. This is
consistent with functional neuroimaging, which shows that left
temporoparietal and inferior prefrontal regions are involved in
the control or selection mechanisms that underpin a variety of
cognitive processes including semantic cognition (23, 25). Sec-
ond, the failure to find anterior temporal lobe activation in
semantic tasks reflects, at least in part, technical limitations of
fMRI. Field inhomogeneities around air-filled cavities lead to
signal dropout and distortions that are particularly pronounced
in orbitofrontal cortex and the inferior and polar aspects of the
temporal lobes (18, 22). Functional neuroimaging that utilizes
PET (which does not suffer from the same problems) does detect
semantically related activation in the anterior temporal lobes,
even when the same experiment conducted in fMRI does not
(22). Because of the preeminence of fMRI in cognitive neuro-
science, however, the potentially central importance of the ATL
within a network of regions that support semantic cognition can
be overlooked (18, 26). Third, results from the outcome of
epilepsy-related resections are complicated by two factors: (i)
longstanding epilepsy might lead to changes in neural organi-
matter connectivity and neurotransmitter function are signifi-
cantly altered in this condition (27, 28); and (ii) this procedure
is unilateral, whereas SD patients have bilateral temporal lobe
atrophy. Other neurological disorders, such as herpes simplex
virus encephalitis, do produce semantic impairment when dam-
age affects the same bilateral temporal lobe regions as semantic
dementia (9, 29).
At the present time there is considerable debate in the
literature about the putative role of different brain regions in
semantic cognition, with strong advocates for the importance of
one brain region over another (18–20). Rather than arguing for
model: an overview of all of these neuropsychological and
neuroimaging studies suggests that semantic cognition is sup-
ported by a three-part neural network made up of the left
prefrontal cortex, the temporoparietal junction, and the tempo-
ral poles bilaterally (7). Although there is convergent evidence
for the involvement of the first two regions, the argument for the
involvement of the temporal poles rests heavily on the SD results
(18). Although the atrophy and hypometabolism are remarkably
circumscribed in this condition (2), it is always possible that the
semantic impairment actually results from damage or infiltration
of pathology in regions beyond those maximally damaged in SD
(19, 20). Accordingly, it is imperative to derive convergent
evidence from neurologically intact participants that the tem-
poral poles are critical regions for semantic memory. We
achieved this aim by utilizing a novel application of repetitive
transcranial magnetic stimulation (rTMS) to induce a ‘‘virtual
lesion’’ in neurologically intact participants. TMS is a well
established noninvasive technique that generates magnetic
pulses over the scalp, inducing electrical activation in a highly
specific area of underlying cortex. A long train of low-frequency
rTMS temporarily suppresses neural processing and disrupts
behavioral tasks that rely on this cortical region. Although the
use of rTMS to probe the function of the temporal poles is new,
TMS has been used to probe other regions and their role in
semantic processing. Consistent with the aphasic and fMRI data
reviewed above, these studies have shown that semantic deci-
sions are slowed after stimulation of the left inferior prefrontal
cortex (and particularly after stimulating the pars orbitalis), and
picture–word verification is slowed after stimulation of left
Wernicke’s area (30, 31).
There are two basic, experimental designs for TMS studies:
the more common ‘‘control site’’ method and the ‘‘control task’’
method (32). If one is interested in testing the neuroanatomical
specificity of a region, then the control site method is most
appropriate. Alternatively, if one is interested in the function of
a specific region (as we are), then the control task method is
more helpful in that one can start to gauge which range of
activities/function the target region is involved in. As noted
above, we already know that semantic cognition is not uniquely
localized to the ATL. Instead, what is controversial is that there
is a role of ATL in semantic cognition. Thus, in designing our
experiment, the focus was to probe the range of functions
supported by the ATL by using the control task method in which
performance on semantic tasks was compared with equally
demanding nonsemantic processes.
The aim of the present study was to test whether rTMS over
the ATL in neurologically intact participants led to a behavioral
‘‘symptom’’ pattern like that observed in SD. Specifically, we
tested whether this produced a combined impairment to com-
prehension and picture naming yet left performance on equally
demanding nonsemantic cognitive tasks (number matching and
Two separate experiments were conducted. In one, participants
named pictures of objects at the basic (e.g., boat, bird) and
specific (e.g., yacht, robin) levels (see Materials and Methods).
They also named six-digit numbers (e.g., 524,673) as a control
task. In the second investigation, participants completed a timed
synonym judgment task and their performance was compared
with a number matching task. We picked these tasks for two
reasons: in our pilot studies we have found that rTMS effects are
patient data (reduced accuracy). Ultra-mild SD gives us clues as
to which types of task are the most sensitive to semantic decline.
In addition to an early drop in accuracy on the same synonym
judgment used here, the patients also have an early and pro-
nounced semantically driven anomia—especially if required to
name at the most specific level (5, 9). Number-related tasks
provide an appropriate cognitive control condition in that
number quantity knowledge is thought to rely on the left parietal
regions and is preserved in SD (34, 35). Furthermore, these tasks
can be configured so that they are as cognitively demanding as
picture naming or word comprehension and thus ensure that any
results on the semantically related tasks cannot be ascribed to
task difficulty alone. To target the temporal pole accurately, this
region was identified individually for each participant by coreg-
istering scalp coordinates with structural MRI scans. In both
experiments we stimulated the same region, 10 mm posterior to
the tip of the temporal pole along the middle temporal gyrus in
the left hemisphere.
The results from the object and number naming tasks are
shown in Fig. 1A. Statistical analysis of reaction times yielded a
main effect of task [F(2,10) ? 34.39, P ? 0.001], as well as a main
effect of TMS condition [F(1,10) ? 5.10, P ? 0.05]. Crucially,
there was a significant interaction between TMS and the naming
task [F(2,10) ? 4.79, P ? 0.03]. Bonferroni tests revealed that
TMS significantly slowed naming responses in the specific-level
task [t(10) ? 3.87, P ? 0.02] but not in the other two tasks
[t(10) ? 1.5, not significant]. In addition, specific-level naming
was slower than both basic-level naming and number naming
[Bonferroni t(10) ? 5.75, P ? 0.001]. A separate analysis of
response accuracy yielded a main effect of the naming task
[F(2,10) ? 11.97, P ? 0.001]. This reflected poorer accuracy in
specific-level naming than basic-level or number naming [13%
errors in specific naming vs. 2% in basic-level naming or 4% in
number naming: t(10) ? 3.75, P ? 0.03].
www.pnas.org?cgi?doi?10.1073?pnas.0707383104Pobric et al.
The results for synonym vs. number judgments are shown in
Fig. 1B. Statistical analysis confirmed that there was a differen-
19.1, P ? 0.002]. Despite being the harder and thus slower task,
number judgment was completely unaffected by temporal pole
stimulation [t(9) ? ?1.08, P ? 0.31] whereas semantic decision
times were slowed, on average, by 9.9% [t(9) ? 7.58, P ? 0.001].
Like the naming tasks, the TMS effect was carried entirely in
speed rather than accuracy. Errors rates were low. Participants
made more errors to the number than synonym judgment task
was no effect of TMS and no interaction (both F ? 1).
These findings demonstrate that the ATL plays a necessary role
in semantic cognition in healthy participants. A temporary
virtual lesion induced by low-frequency rTMS over left ATL
significantly increased naming latencies for a specific-level nam-
same region significantly slows synonym judgment times but not
number quantity decisions. The results of this study mirror the
core features of semantic dementia (SD) patients, who show a
progressive and highly selective deterioration of semantic mem-
ory in the context of focal atrophy of the ATL (2, 5). Like the
patients, rTMS over the ATL produces a combination of im-
paired comprehension and a resultant naming impairment but
does so in a selective manner. The parallelism between the
patient data and these novel rTMS results is consistent with the
notion that the anterior temporal lobe region integrates all types
of verbal and nonverbal information into an amodal represen-
tation that allows all information associated with a concept to be
activated from any particular, single input modality and also
allows for appropriate generalizations from one concept to
Studies of various patient groups and functional neuroimaging
in normal participants have consistently demonstrated a critical
role of left prefrontal and temporoparietal regions in semantic
cognition (21, 31, 36). When data from SD patients are com-
bined with convergent results from this temporal pole rTMS
study, then it becomes clear that semantic cognition is actually
supported by a three-region neural network: left prefrontal,
temporoparietal, and bilateral anterior temporal regions. Pre-
vious comparative neuropsychological studies suggest that there
is a division of labor across these areas such that core semantic
representations are reliant on the anterior temporal lobes
whereas semantic control—like other forms of executive
control—is reliant on prefrontal–temporoparietal circuitry (23,
25). In the undamaged system these regions interact to support
flexible, temporally extended semantic behavior (semantic cog-
nition). With impairment to the anterior temporal lobe, core
semantic representations become degraded and patients are
unable to activate all of the information associated with a
concept (6, 7, 9). Multimodal comprehension deficits can also
systems. In these circumstances the patients are unable to
reliably shape or control the aspects of meaning that are relevant
for the task at hand or are critical at specific moments during
temporally extended tasks (7).
In this study rTMS significantly slowed specific-level naming,
but there was only a weak, nonsignificant trend to slow basic-
level naming. The ATL atrophy of semantic dementia also
induces a specificity effect in both comprehension and naming
(6, 8, 9). Likewise, a recent PET functional imaging study
required to verify picture–name pairings at the specific level
(37). There are two possible explanations for this effect. The first
is that specific-level conceptual differences are simply more
demanding for the normal semantic system and thus are the first
to exhibit effects of brain disease, show greater neural activation,
or, in this study, the effects of rTMS (6, 9). An alternative
explanation is that there is a basic-to-specific representational
gradient along the temporal lobe such that the anterior regions
are specialized for specific concepts (38). Although this alter-
native hypothesis is consistent with the rTMS results presented
here, other neuroimaging and patient data would seem to argue
against it. Although SD patients show worse performance for
specific-level than basic-level distinctions, their performance on
basic-level concepts is also impaired, suggesting a graded dif-
ference underpinning the specificity effect (6). In addition, a
number of neuroimaging studies have found that anterior tem-
poral lobe regions are activated in normal participants by speech
production or comprehension tasks involving basic-level con-
cepts (22, 39). Such findings are consistent with the notion that
ATL regions form an amodal semantic representation system
and that performance differences reflect systematic variations in
the formation and thus representation of these concepts (6, 26).
Materials and Methods
Participants. Twelve participants took part in the naming tasks
(seven females; mean age ? 21.7, SD ? 4.05). Ten participants
took part in the synonym and number judgment tasks (six
females; mean age ? 22.3 years, SD ? 4.82). All participants
were native English speakers and strongly right-handed, yield-
Level/type of naming
/ ( s e
i t g
/ ( e
i t n
o i s i c
judgments (B). Each bar represents mean reaction time for that condition.
Error bars represent the standard error for each mean. A summarizes data
synonym and number judgment tasks.
Effect of rTMS to the left temporal pole on naming (A) and synonym
Pobric et al. PNAS ?
December 11, 2007 ?
vol. 104 ?
no. 50 ?
ing a laterality quotient of at least ?90 on the Edinburgh
Handedness Inventory (40). They were free from any history
of neurological disease or mental illness, and they were not on
any medication. All had normal or corrected-to-normal vision.
The experiments were reviewed and approved by the local
ethics board (Central Office for Research Ethics Committees
Design. A within-participant factorial design was used in both
experiments, with TMS (no stimulation vs. temporal pole stim-
ulation) and task (basic vs. specific vs. number naming or
synonym vs. number judgment) as the two within-participant
factors. The study utilized rTMS using the virtual lesion method
in which the train of rTMS is delivered offline (without a
concurrent behavioral task) and then behavioral performance is
probed during the temporary refractory period and compared
with performance on the same task outside this refractory
window. In pilot studies we found that semantic decision times
were suppressed for ?20 min after 10 min of 1-Hz rTMS. We
also found that rTMS and the associated novel experience,
irrespective of site of stimulation, is highly alerting for partici-
pants. As a consequence there is a nonspecific speeding of
reaction times (on all tasks). Accordingly, the study was designed
to deconfound order and the specific TMS effect. Half of the
participants produced their ‘‘baseline,’’ no-TMS data before
rTMS was applied. The other half provided their baseline at least
30 min or more after the end of rTMS (by which time our pilot
studies indicate that no behavioral effect remains).
Stimuli and Procedure. Naming tasks. A total of 128 picture stimuli
and 64 number stimuli were used. To allow for direct compar-
isons with SD data, these stimuli were taken from existing
64 line drawings were taken from the Snodgrass and Vanderwart
set (41), covering various different categories (animals, birds,
fruit, household items, tools, and vehicles) and are the same
items as those used to assess basic naming function in some
studies of semantic dementia (4). For the specific naming task,
there were 64 color photographs drawn from the same semantic
categories. Colored pictures were required for this condition,
rather than line drawings, to identify the target, specific concept
uniquely. There is no evidence to suggest that the type of
material affects performance in semantic dementia, and, if
anything, one might expect colored pictures to improve perfor-
mance on the specific level condition, thus working against the
specificity effects found here (33). For these specific-level items
the participants were asked to provide a specific-level name (i.e.,
swan, poodle, mini, hovercraft). In pilot tests, these items had
?95% name agreement at this specific level. For the number
naming task, participants provided English names for six-digit
hundred and sixty-five’’). In pilot studies we had found that these
longer numbers provided longer naming latencies that matched
the typical basic-level naming speeds. Each group of items was
split into two sets matched for name frequency and age of
acquisition, one set being used in the baseline/no-TMS condition
and the other immediately after the rTMS (see Design). The two
sets were counterbalanced across participants. A PC running
E-Prime software (Psychology Software Tools) presented the
stimuli and recorded the responses. The participants sat in front
of a 15-inch monitor and were instructed to name the stimuli as
quickly and accurately as possible. Participants performed all
three naming tasks in each experimental session. The order of
the tasks was counterbalanced across subjects. There were eight
practice trials for each stimulus set followed by 32 experimental
trials in a random order. A fixation point appeared on the screen
to signal the start of each trial. The participant then pressed a
space bar to display the stimulus to be named. Stimuli were
presented until the response was given and were followed by a
blank screen interval of 500 ms. Verbal responses were recorded
by using a microphone placed in front of each participant. The
latency of each response was recorded by the computer, and the
accuracy was checked offline by listening to audio recordings.
Response times that did not fall within 2 SD of the mean for each
participant in each condition were discarded. This resulted in the
removal of ?3.7% of all responses. One participant made a large
number of errors and was excluded from the analysis.
Judgment tasks. The synonym judgment task was based on a
neuropsychological assessment that we have developed to test
verbal comprehension in SD and other aphasic patient groups.
The 96 trials from the clinical test were augmented with addi-
tional trials to provide enough trials for the TMS and no-TMS
versions. The final experiment includes two versions containing
72 trials each (144 in total), matched for the imageability and
frequency of the words. Each trial contains four words: a probe
word (e.g., rogue), the target choice (e.g., scoundrel), and two
unrelated choices (e.g., polka and gasket). The number task also
judgment task: a probe number was presented at the top of the
screen, and underneath three number choices were given. Par-
ticipants were required to pick which of the three was closest in
value. In pilot studies we found that, by using double-digit
numbers, the resultant number judgment times were typically
slightly slower and less accurate than the synonym judgment
tasks (see Results). Accordingly, any specific effects of temporal
pole rTMS on synonym judgment could not be due to task
A PC running E-Prime software allowed the presentation of
stimuli and recording of the responses. Participants performed
two synonym and number judgment tasks per experimental
session (one within and one outside the rTMS-induced refrac-
tory period; see above). The experiment began with a practice
block of six trials for each stimulus set. Experimental trials were
presented in a random order in four blocks of 72 trials (two
blocks of the same task). A fixation point appeared on the screen
to signal the start of each trial. The participant then pressed a
space bar, which advanced the experiment on to the next
stimulus. Stimuli (words, numbers) were presented until re-
were asked to indicate the synonym of the probe word, or which
number was closest in magnitude to the probe number, by
pressing with the right hand one of three designated keys on a
keyboard. The two versions of the tasks were counterbalanced
across participants. As noted above, whether the non-TMS
session was conducted before or after (at least 30 min) the
TMS was counterbalanced across participants to deconfound
TMS and order effects.
rTMS Procedure. Focal magnetic stimulation was delivered by
stimulator (Magstim). For every subject and in each session,
motor threshold (MT) was determined before the experiment as
a visible twitch in the relaxed contralateral abductor pollicis
brevis muscle. Stimulation was set at 120% of MT corresponding
to an average intensity of 67% ? 6.88 (mean ? SD) of maximum
stimulator output for the participants in the naming study and
77 ? 6.88% for those in the matching experiment. Structural
T1-weighted MRI scans were obtained for each subject to guide
the positioning of the TMS coil. Scalp position was coregistered
with the underlying cortical surface for each individual by using
an Ascension minibird magnetic tracking system and MRIreg
software (www.mricro.com/mrireg.html). Eight fiducial markers
(oil capsules attached to nasion, vertex, inion, tip of the nose,
left/right mastoids, and left/right tragus during scanning) were
as the region 10 mm posterior from the tip of the left temporal
www.pnas.org?cgi?doi?10.1073?pnas.0707383104Pobric et al.
pole along the middle temporal gyrus. This point was used in Download full-text
each participant as an anatomical landmark for the temporal
pole (TP). The location of the TP was identified on each
participant, and the scalp location directly above this site was
for the ATL in standard space were (?53, 4, ?32). At this site
regions (including the hand area), and rTMS effects may be
somewhat weaker as a consequence (although we have found
that this distance varies considerably from individual to individ-
ual). The coil was held securely over the site to be stimulated and
oriented such that the maximal induced current flowed approx-
imately in the anterolateral direction along the middle temporal
gyrus. Particular care was taken in the placing of the coil because
TMS over ATL is more uncomfortable than over occipital or
parietal areas. Participants received TMS stimulation for 10 min
(1 Hz for 600 s at 120% MT).
The study was supported by a Research Councils UK fellowship,
Wellcome Project Grant 078734/Z/05/Z, and Medical Research Council
Program Grant G0501632.
1. Hodges JR, Patterson K, Oxbury S, Funnell E (1992) Brain 115:1783–1806.
2. Nestor PJ, Fryer TD, Hodges JR (2006) NeuroImage 30:1010–1020.
3. Galton CJ, Patterson K, Graham KS, Lambon Ralph MA, Williams G, Antoun
N, Sahakian BJ, Hodges JR (2001) Neurology 57:216–225.
4. Bozeat S, Lambon Ralph MA, Patterson K, Garrard P, Hodges JR (2000)
J Cognit Neurosci 13:341–356.
6. Rogers TT, Lambon Ralph MA, Garrard P, Bozeat S, McClelland JL, Hodges
JR, Patterson K (2004) Psychol Rev 111:205–235.
7. Jefferies E, Lambon Ralph MA (2006) Brain 129:2132–2147.
8. Warrington EK (1975) Q J Exp Psychol 27:635–657.
9. Lambon Ralph MA, Lowe C, Rogers TT (2007) Brain 130:1127–1137.
10. Coccia M, Bartolini M, Luzzi S, Provinciali L, Lambon Ralph MA (2004)
Cognit Neuropsychol 21:513–527.
11. Luzzi S, Snowden JS, Neary D, Coccia M, Provinciali L, Lambon Ralph MA
(2007) Neuropsychologia 45:1823–1831.
12. Lambon Ralph MA, Graham KS, Patterson K, Hodges JR (1999) Brain Lang
13. Bozeat S, Lambon Ralph MA, Graham KS, Patterson K, Wilkin H, Rowland
J, Rogers TT, Hodges JR (2003) Cognit Neuropsychol 20:27–47.
14. Bozeat S, Lambon Ralph MA, Patterson K, Hodges JR (2002) Cognit Affective
Behav Neurosci 2:236–251.
15. Bozeat S, Lambon Ralph MA, Patterson K, Garrard P, Hodges JR (2000)
16. Gloor P (1997) The Temporal Lobe and the Limbic System (Oxford Univ Press,
17. Damasio H, Grabowski TJ, Tranel D, Hichwa RD, Damasio AR (1996) Nature
18. Wise R (2003) Br Med Bull 65:95–119.
19. Hickok G, Poeppel D (2007) Nat Rev Neurosci 8:393–402.
20. Martin A (2007) Annu Rev Psychol 58:25–45.
21. Berthier ML (2001) Aphasiology 15:99–130.
22. Devlin JT, Russell RP, Davis MH, Price CJ, Wilson J, Moss HE, Matthews PM,
Tyler LK (2000) NeuroImage 11:589–600.
23. Garavan H, Ross TJ, Li SJ, Stein EA (2000) Cereb Cortex 10:585–592.
24. Hermann B, Davies K, Foley K, Bell B (1999) Epilepsia 40:1070–1076.
25. Peers PV, Ludwig CJH, Rorden C, Cusack R, Bonfiglioli C, Bundesen C,
Driver J, Antoun N, Duncan J (2005) Cereb Cortex 15:1469–1484.
26. Rogers TT, McClelland JL (2004) Semantic Cognition: A Parallel Distributed
Processing Approach (MIT Press, Cambridge, MA).
27. Powell HWR, Parker GJM, Alexander DC, Symms MR, Boulby PA, Wheeler-
Kingshott CAM, Barker GJ, Koepp MJ, Duncan JS (2007) NeuroImage
28. Hammers A, Koepp MJ, Richardson MP, Hurlemann R, Brooks DJ, Duncan
JS (2003) Brain 126:1300–1318.
29. Noppeney U, Patterson K, Tyler LK, Moss H, Stamatakis EA, Bright P,
Mummery C, Price CJ (2007) Brain 130:1138–1147.
30. Knecht S, Floel A, Drager B, Breitenstein C, Sommer J, Henningsen H,
Ringelstein EB, Pascual-Leone A (2002) Nat Neurosci 5:695–699.
31. Devlin JT, Matthews PM, Rushworth MFS (2003) J Cognit Neurosci 15:71–84.
32. Jahanshahi M, Rothwell J (2000) Exp Brain Res 131:1–9.
33. Moore CJ, Price CJ (1999) Brain 122:943–962.
34. Butterworth B, Cappelletti M, Kopelman M (2001) Nat Neurosci 4:784–786.
35. Jefferies E, Patterson K, Jones RW, Bateman D, Lambon Ralph MA (2004)
36. Thompson-Schill SL, Desposito M, Aguirre GK, Farah MJ (1997) Proc Natl
Acad Sci USA 94:14792–14797.
37. Rogers TT, Hocking J, Noppeney U, Mechelli A, Gorno-Tempini ML,
Patterson K, Price CJ (2006) Cognit Affective Behav Neurosci 6:201–213.
38. Martin A, Chao LL (2001) Curr Opin Neurobiol 11:194–201.
39. Mummery CJ, Patterson K, Hodges JR, Wise RJS (1996) Proc R Soc London
Ser B 263:989–995.
40. Oldfield RC (1971) Neuropsychologia 9:97–113.
41. Snodgrass JG, Vanderwart M (1980)J Exp Psychol Hum Learn Mem 6:174–215.
Pobric et al. PNAS ?
December 11, 2007 ?
vol. 104 ?
no. 50 ?