I've been working on my Psych paper for most of the day and this is the resulting introduction and beginning to the methods. I still need to work out some of the experimental details, but they'll involve localizing the FFA, programming the TMS to knock it out, and behavioral data. I'll probably try to make some images with MRIcro that'll be a picture of a brain in comparison to another brain and I'll fudge around in photoshop to make "stimuli". Other than that, it looks good except for length. I can't go over five pages, so that's a really big concern right now.
Introduction
Part of the ventral temporal cortex, the FFA is known through previous function magnetic resonance imaging (fMRI) studies to be highly sensitive to the presentation of faces (Kanwisher, et al. 1997). Neuropsychological evidence suggests that the presence of a functional FFA is required to discriminate and identify faces. Children suffering from autism often show difficulty recognizing faces due to dysfunction in the FFA (Pierce, et al., 2001). Those autistic children have little to no activity in the fusiform gyrus compared to a normal subject and develop regions highly sensitive to faces outside of the expected area of the fusiform gyrus. Similarly, patients with lesions to the fusiform gyrus often suffer from prosopagnosia (the inability to perceive faces). In many cases, prosopagnosia is coupled with visual agnosia as lesions are usually large and impairs much of the ventral visual stream, but, in some cases, only facial recognition is impaired without any deficits in object recognition due to restricted damage to the fusiform gyrus (e.g., Wada & Yamamoto, 2001). Additionally, visual agnosia can occur without also leading to prosopagnosia (e.g., Moscovitch, et al. 1997, as cited in Kanwisher & Yovel, 2006). Coupled with fMRI data that shows the FFA as a highly activated region, the neuropsychological data provides strong evidence to suspect the FFA as a vital component of the facial recognition process.
Two models currently exist to describe the activity of the fusiform face area (FFA), but several debates question the extent to which each model fully encompasses its actual function. In one model, the FFA acts as a face-specific domain, responsible solely for the face perception (Kanwisher & Yovel, 2006). In this model, faces are perceived and processed on a holistic level, recruiting this specific region to discriminate and individuate between perceived faces. The opposing model proposes that perceived faces do not specifically and exclusively recruit the FFA to process information, but only follows a visual processing pathway used to discriminate between objects of high familiarity and expertise (Tarr & Gauthier, 2000). In particular, experts in distinguishing between types of cars and experts in distinguishing between types of birds showed increased activity in the FFA over other novel objects (Gauthier, et al., 2000). This increased activity in the defined FFA region spanned other forms of expert recognition, even newly trained expertise in novel geometric shapes called “greebles” (Gauthier, et al., 1999).
The two models are not easily discriminated through correlative techniques such as fMRI; however, recent studies utilizing high-resolution fMRI (HR-fMRI) have demonstrated unique face specific activations corresponding single neurons that show high sensitivity to faces, but low sensitivity to other non-face objects (Grill-Spector, et al., 2006). The study also provides an explanation for many of the findings that form the foundation of the expertise hypothesis. Grill-Spector identifies several regions highly sensitive to other non-face stimuli, but not to faces, that are heterogeneously interspersed within the FFA region. These object-sensitive regions are small, often single voxels, observed only with the increased spatial detail of HR-fMRI; but, with standard fMRI techniques, are spatially averaged along with nearby face-sensitive regions to produce a region that appears sensitive to both faces and other non-face objects.
The limitations of fMRI correlative studies prevent studies from directly asserting causation between observed activity and unobserved function. The solution to explicitly demonstrate causation lies in transcranial magnetic stimulation (TMS) of the region in question. TMS provides a method to specifically target certain regions of the brain, selectively exciting or inhibiting their function by pulsing a magnetic field through the cranium (Hallet, 2000). The magnetic field creates a current that opposes or stimulates the current flow in the neurons. This selective impairment allows targeted research into the function of a specific region in the context of a neural network.
Neuropsychological data certainly suggests that the fusiform gyrus, in general, and the FFA, in particular, plays a vital role in facial recognition, but this study attempts to further solidify that suggestion. By examining the effects TMS of the FFA in subjects with high expertise in birds and cars, this study hopes to separate facial recognition from expert recognition as if the FFA were to act as a domain-specific region. In that case, TMS of the FFA should induce temporary prosopagnosia without impairing subjects’ expert recognition in their field of expertise.
Methods
Subjects
Subjects were recruited in three categories: car-experts, bird-experts, and controls. 13 car-experts (4 females, age 30-48), 17 bird-experts (10 females, age 28-47), and 21 control subjects (9 females, age 19-49) were recruited for the study. All car and bird experts had significant training in their respective expertises (average car expertise - 15.6 years; average bird expertise – 13.2 years) and were tested prior to the study to verify their expertise. Controls were recruited locally through community advertisements and had no more knowledge of birds or cars than expected in the general population. All subjects were provided monetary compensation for their time. Subjects gave written informed consent prior to participating in the study.
Apparatus
Subjects were scanned on a Varian 4-Tesla INOVA scanner at the Henry H. Wheeler, Jr. Brain Imaging Center at the University of California, Berkeley in Berkeley, CA. TMS was performed using a MR-safe non-ferromagnetic coil synchronized to fire between MRI image captures. A dummy coil made of non-ferromagnetic materials was substituted in subjects not subjected to TMS in order to control for effects due to the presence of the coil.
