вход по аккаунту


Differential neural activity in the human temporal lobe evoked by faces of family members and friends.

код для вставкиСкачать
Dlfferential Neural Activity in the
Human Temporal Lobe Evoked by Faces
of F d y Members and Friends
M. Seeck, MD,* N. Mainwaring,? J. Ives,? H. Blume, M D , PhDJ D. Dubuisson, MD, PhD,$
R. Cosgrove, MD,§ M.-M. Mesulam, M D , t and D. L. Schomer, M D t
I n 6 patients, depth electrodes revealed differential evoked responses to familiar versus novel faces. These differential
responses were obtained in the amygdala, hippocampus, and temporal neocortex but not in the dorsolateral frontal
or cingulate cortex. T h e limbic and temporal structures that differentiated novel from familiar faces did not respond
differentially to variations in luminance. Limbic structures and temporal cortex thus appear to participate in face
recognition and in encoding the familiarity of visual experiences.
Seeck M, Mainwaring N, Ives J, Blume H, Dubuisson D, Cosgrove R, Mesulam M-M,
Schomer D. Differential neural activity in the human temporal lobe evoked
by faces of family members and friends Ann Neurol 1993;34 369-372
Components of the temporal lobe such as the amygdala, hippocampus, and association cortex play a major
role in memory functions, emotional labeling, and
complex visual processing. T h e role of the temporal
lobe in face recognition has been emphasized in numerous animal and human studies {l-141. Physiological recordings in monkeys show that neurons in the
inferotemporal cortex and amygdala are tuned selectively to faces and their expressions [ 1-6). In humans,
case studies have consistently shown that bilateral temporooccipital lesions can lead to profound prosopagnosia, that is, the inability to identify faces, even those
that are familiar r7,81.
In order to further investigate the neura1 correlates
of face recognition, w e presented photographs of familiar and novel faces to 6 patients in whom intracranial
electrodes had been inserted for diagnostic purposes.
O u r purpose was to determine whether familiar faces
would evoke responses that were different from those
evoked by unfamiliar faces and whether the distribution of such differential response patterns would display anatomical specificity.
In 6 patients with complex partial epilepsy refractory to medication, depth electrodes were implanted bilaterally in the
brain in order to identify a localized or focal area of epileptogenesis and consider its surgical removal. Contact sites included the amygdala, hippocampus, parahippocampal gyrus,
From the 'Department of Neurology, Ludwig-Maximilian Universiw, Munich, Germany; the tDcpartmenr of Neurology, $Division
of Neurosurgery, Beth Israel Hospital, Harvard University, Boston,
MA; and the sDepartmcnt Of Neurosurgery,
Of Virginia,
Charlottesville, VA.
temporal neocortical areas, the anterior cingulate gyrus, and
orbitofrontal and dorsofrontal areas. Each depth electrode
had six to eight regularly spaced recording sites (Fig 1). The
electroencephalogram (EEG) was obtained after amplification, digitization (200 Hz), and filtering (bandpass, 0.1-70.0
Hz), and was stored and analyzed off-line (prestimulus baseline, 100 msec). Age, handedness, and information about
seizures are included in Table 1. Magnetic resonance imaging
did not show structural lesions in any of the 6 patients. Informed consent was obtained prior to testing.
The face stimuli we employed were either familiar or novel
(n = 172) to each patient. The set of familiar faces was
individualized for each patient. It consisted of 35 to 42 faces
of close family members and of friends and neighbors who
were known to the patient for decades. The set of novel
faces was obtained from the identification office of an out-oftown bank. All face photographs were digitized and shown
as black-and-white pictures on a high-resolution computer
screen. The faces were presented in sequential pairs of two,
with a 500-msec exposure for each and an interval of 150
msec between members of each pair. A randomized interval
of 1,800 to 2,200 msec was given until the next pair of
pictures was shown. All faces were smiling or showing a neutral expression in a frontal view. Each familiar and unfamiliar
face was presented two to four times (familiar faces, 2.4
0.7 standard deviation [SD}; unfamiliar faces, 2.2 & 0.4 SD;
difference not significant). After artifact rejection iinterictal
epileptiform activity, etc.) the remaining visual evoked responses were averaged. A total of 35 to 40 sweeps related
to familiar faces was recorded for each patient, compared to
approximately 160 sweeps during exposure to unfamiliar
Reccived Ian 28, 1993, and in revised form Aur 16. Acceuted for
publication Apr 19, 1993.
Address correspondence to Dr Schomer, Clinical Neurophysiology,
Beth Israel Hospital, 330 Brookline Avenue-GZ522, Boston, MA
Copyright 0 1993 by the American Neurological Association 369
Fig I . Magnetic resonanie image IMRI) from Patient 1 with
depth electrodes in the left medial temporal lobe. Right-J ided rlrctrodes would be .teen in the next M R I section.
faces. In addition to familiarity,luminance was also varied in
a second set of novel faces.The patients were asked to press
the r g h t (ALT) button on a conventional keyboard if the
face seemed familiar and a right (CRTL) button if this was
not the case.
All patients identified the familiar faces with 100F accuracy, except Patient 6 who missed 4 out of 18 faces. No
novel face was misidentified as familiar. The EEG responses
related to these misidentifications were not included in the
analysis. Our studies were done within the first week of EEG
monitoring (3-7 days after implanting the electrodes) prior
to any significant withdrawal of anticonvulsants.A distinction
between averages for EEG responses related to familiar and
compared to unfamiliar faces was defined as differences
greater than 1 standard error (SE).
In all our patients, prominent category-related differences between familiar and unfamiliar faces were found
in medial temporal lobe recordings. The findings are
summarized in Table 2. The differences of potentials
evokcd by familiar and unfamiliar faces consist of different latencies of the same peak (Patient 1; Fig 2A),
peaks evoked by familiar faces but not by unfamiliar
faces (Patients 3 , 5 , and 6; Fig 2B), or diverging evoked
potential patterns (Patients 2, 4 , and 5 ; Fig 2C). The
right amygdala was the one structure most frequently
(in 4 out of 6 patients) associated with differential
evoked responses.
In 5 patients (Patients 1, 5 , and 6 on the right; Patients 2 and 3 on the left) marked differences were
also found in the cortex of the superior and/or inferior
temporal sulcus. In all cases, the major differences
were found after 150 to 200 msec. Such a latency is
consistent with the onset of higher-order perceptual,
mnemonic, and emotional processing. Differential responses to familiar as opposed to novel faces were not
recorded in other recording sites such as the dorsolateral frontal lobe or cingulate gyrus. Orbitofrontal recordings were obtained in 3 patients and differential
responses were noted in only 1. Variations in luminance did not evoke differential responses in the amygdala, hippocampus, or the ocher temporal lobe contact
sites (Fig 2D).
One or more structures of the temporal lobe in each
patient we tested consistently displayed a differential
response to familiar and unfamiliar faces 150 to 200
msec after stimulus onset. These results show that familiar faces evoke a pattern of neuronal response that
is distinct from that evoked by unfamiliar faces, and
suggest that the temporal lobe of the human brain is
specialized for recognizing experiential attributes such
as familiarity associated with face-related stimuli. The
encoding of “familiarity” by structures of the medial
temporal lobe is consistent with phenomena such as
dejh vu and jamais vu that are frequently reported by
patients with temporal lobe epilepsy.
Table 2 shows that the differential responses to fa-
Table 1 . Patient Characteristics
Age of
Onset (yr)
(Site of Focus)
Left = right temporal
Left temporal
Right occipitotemporal
Right temporal
Lcft temporal
Right > left temporal
Patient No.
Age (yr)
370 Annals of Neurology Vol 34 No 3 September 1993
Table 2. Study Results"
Patient No.
Frontal Cortex
Temporal Cortex
"Plus signs ( + ) indicate structures where there is a difference between averagcs of evoked potentials for familiar and unfamiliar faces. Minus
signs ( - ) indicate an absence of difference between averages of evoked potentials for familiar and unfamiliar faces. No discriminating responses
were found in frontal structures and the cingulate gyrus.
ND = no data available due to technical causes; + + = visual evoked potential averages more than 2 standard errors (SE) apart (SE of the
unfamiliar face evoked potential); + = visual cvoked potential averages between i and 2 SE apart.
-I- '
500 ms
5 0 0 rns
500 ms
Fi g 2. (A-C) Visual wokedpotentzals (VEPs,. The thick lines
indicate familiar faces! and the thin lines, urzfamiliarfaces.
Both are shown with the I standard-emr range. (A) VEPs of
Patient I from the right aniygdala. Familiar faces emked an
earlier negative peak thaiz did the unfaniiliar faces (delta, 2060 mec), as well as a more positive shape of the VEP. Uzflerences betuieen the t w o categories start as early as I S 0 nisec.
(B)V E P of Patient 3 from the lejit hippocampus. In thij-patient,
familiar /aces resulted in a negatioe peak at 120 msec with no
particular ez!oked potential pattern related t o unfamiliar fares.
(C) VEP of Patient J from the right temporal cortex. Both stimulus categories evoked particular responses differing from the
baseline and from each other after 200 ni~ec.(D? VEP of Patient 5 from the same depth contact as in C , showing the eflect
of wivying the ltdminance of novel faccs. The thick line indicates
dark faces (around .3 cdfnz2)and the thin line, bright faces
(around 30 cdim'?. Both are shown with the 1 standurd-error
range (the darkened picture uas achiewd bji proportionately reducing the brightness of the 16 gray levels by 90%). l h e evoked
responses are not .rignz$cuntlj dffeevent.
Seeck et al: Differential Evoked Responses to Familiar Faces
miliar faces were more frequently obtained in the right
than in the left hemisphere. This finding is consistent
with neuropsychological evidence showing that the
right hemisphere is more closely involved in face recognition [l 1, 12). Differential responses to familiar
faces were obtained at multiple sites of the temporal
lobes but not in dorsolateral frontal cortex or the cingulate gyrus. As expected, the temporal lobe sites that
gave differential responses to familiar and novel faces
did not respond differentially to changes of luminance
since this type of discrimination occurs at regions of the
visual cortex that were not covered by our electrode
placements. There was therefore both anatomical and
stimulus specificity in the results that we obtained. The
6 patients did not display uniform patterns of evoked
responses to either novel or familiar faces. Seizureinduced alterations of cortical organization, handedness
and personal peculiarities in recognition style, memory organization, and emotional labeling could account
for some of the interindividual differences shown in
Figure 2.
Microelectrode studies in monkeys have revealed
enhanced neuronal firing related to face recognition in
the amygdala and the cortex of the superior temporal
sulcus [ 3 , 5, 91. A recent magnetoencephalographic
study in humans showed that areas of the middle temporal gyrus are activated by faces but not by other
visual stimuli [lo}. In patients undergoing right-sided
craniotomy, microelectrode studies found increased
neuronal activity in the right anterior and medial
temporal cortex during face-matching tasks I1 11. In
healthy volunteers engaged in a face identification task,
positron emission tomography showed significant bilateral activation of the temporal lobe, predominantly in
the right anteromediai temporal gyrus and the parahippocampal area [12). Human lesion studies suggest that
the area5 critical for facial recognition also extend more
posteriorly into the lingual-fusiform region of the inferomedial temporal lobe, but these areas were not
covered in our study since the electrode placements
were chosen for the investigation of epilepsy, not for
the investigation of face recognition CG, 13, 14).
Tzaken together with our findings, these results show
that the middle, inferior, and medial temporal lobe
372 Annals of Neurology Vol 34
structures of the primate brain contain a widely distributed neural network that plays a critical role in the
identification of faces. Our study is the first to demonstrate that these temporal lobe structures of the human
brain participate in the differential encoding of faces
that are personally familiar to the individual.
1. Ungerleider LG, Mishkin M. Two cortical visual systems. In:
Ingle DJ, Goodale MA, Mansfield KJW, eds. Analysis of visual
behavior. Boston: MIT Press, 1982:549-586
2. Rolls ET. Neurones in the cortex of the temporal lobe and the
amygdala of the monkey with responses selective for faces. Hum
Neurobiol 1984;3:209-222
3. Leonard CM, Rolls ET, Wilson FAW, Baylis GC. Neurons in
the amygdala of monkeys with responses selective for faces.
Dehav Brain Res 1985;15:159-176
4. Perrett DI, Smith PAJ, Potter DD, et al. Neurones responsive
to faces in the temporal cortex: studies of functional organization, sensitivity to identity, and relation to perception. Hum
Neurobiol 1984;3:197-208
5. Perrett D1, Mistlin AJ. Perception of facial characteristics in
monkeys. In: Stebbins WC, Berkley MA, eds. Comparative perception, vol 11: complex signals. New York: Wiley, 1990:187-
6. Brothers L, Ring B, Kling A. Response of neurons in the ma-
No 3 September 1993
caque amygdala to complex social stimuli. Behav Brain Kes
1990;41: 199-2 13
Damasio AR, Damasio H, Van Hoesen GW. Prosopagnosia:
anatomic basis and behavioral mechanisms. Neurology 1982;
Landis T, Regard M, Bliesde A, Kleihus P. Prosopagnosia ,and
agnosia for noncanonical views. Brain 1988;111:1287-1297
Baylis GC, Rolls ET, Leonard CM. Functional subdivisions of
the temporal lobe neocortex. J Neurosci 1987;7:330-342
Lu ST, Hamalainen h l S , Hari R, et al. Seeing faces activates
three separate areas outside the occipital visual cortex in man.
Neuroscience 1991;43:287-290
Ojemann JG, Ojemann GA, Lettich E. h'euronal activity related
to faces and matching in human right nondominant temporal
cortex. Brain 1972;115:1-13
Sergent J, Ohta S, MacDonald B. Functional neuroanatomv of
face and object processing. Brain 1992;115:15-36
Haxby JV, Grady CL, Horowitz B, et al. Dissociation of object
and spatial visual processing pathways in extrastriate cortex. Proc
Natl Acad Sci USA 1991;88:1621-1625
Mesulam MM. Higher visual functions of the cerebral cortex
and their disruption in clinical practice. In: Albert DM, Jakobiec
FA, eds. The principles and practice of ophthalmology: the Harvdrd system. Philadelphia: Saunders (in press)
Без категории
Размер файла
400 Кб
facer, family, neural, differential, activity, friends, members, temporal, human, evoked, lobel
Пожаловаться на содержимое документа