User:Lu.roberts1991/sandbox

The N170 is a component of the event-related potential (ERP) linked with the processing of faces. Event-related potentials are recorded using electroencephalography (EEG) and are linked to specific processes in the brain, in research ERPs are recorded after the onset of the stimulus. The peaks can either be positive or negative; reflected in the letter at the beginning of their name. The time at which they occur in milliseconds (ms) after stimulus onset is reflected in the number in the name. Whether the peak is positive or negative has no significance as it depends on the baseline electrical activity, it is the amplitude of the peak that is of interest. The N170 component onset after stimulus presentation and duration occurs at about 130 to 200 ms, and its amplitude shows more negativity after the presentation of faces than after other stimuli. The negativity and onset is reflected in the name N170. The sensitivity to faces is most obvious over occipito-temporal electrode sites ; the source of this effect is thought to be the fusiform and inferior-temporal gyri. This is consistent with other research that pinpoints these areas as important in face processing. The N170 is also lateralised to the right hemisphere.

History
The first face sensitive component to be discovered was not in fact the N170 but rather the Vertex-Positive Potential (VPP). The onset of this component is between 140 and 180 ms and peaked at fronto-central sites. This was first reported by Bötzel and Grusser in 1989 and repeated by Jeffreys in the same year. Bötzel and Grusser presented line drawings and black and white photographs of faces, chairs and trees to participants. They found the VPP onset to be approximately 150 ms after stimulus onset, and that it had a higher amplitude in response to faces than the other stimuli categories. They concluded that this component, given its location, was originating from the limbic system.

Research in the 1990's sparked debate regarding the VPP, as more studies began to find a negative component, over the occipito-temporal areas. One such study was conducted by Bötzel and colleagues 1995 who presented black and white images of faces, leaves and flowers to their participants. Although they found the VPP at 180 ms for all images with a higher amplitude for faces, they also reported a negative peak at 175 ms in the posterior temporal electrodes, which like the VPP showed higher amplitude after the presentation of faces. They argue that because the previous studies that looked at the VPP used a mastoid reference electrode rather than an electrode from the scalp as they did that these two components may in fact be part of the same face specific process. For an explanation of reference electrodes see the method section of the EEG page.

More recently, in 2005, Joyce and Rossion conducted a study to discover whether or not the N170 and the VPP are a part of the same process and that the differences in the literature are a result of reference location. They used recordings from the left mastoid, right mastoid, earlobes, nose, right sternoclavicular junction and vertibrae C7 to reference the EEG data, which they recorded while presenting faces, cars and words. They concluded that these two components are 'two faces of the same brain generators'.

Bentin and colleagues in 1996 were the first to give the negative component in the posterior temporal areas the name of N170. The participants were given a 'visual target detection' task. The targets were embedded in other stimuli, these included upright, distorted and inverted faces, face components, animal faces and non-face stimuli. The N170 was evoked 172 ms after the onset of upright faces, distorted faces evoked an N170 similar in amplitude. The N170 was largest in those electrodes over the posterior temporal areas. They also found that the N170 was lateralised to the right hemisphere. For inverted faces the N170 was elicited later than upright faces and when eyes were presented alone they evoked a significantly larger N170 than when the whole face was presented. The other sections of the face, lips and noses, however elicited a smaller negative component that appeared approximately 50 ms later than the N170. They reported that there was no N170 evoked after the presentation of all the non-face and animal face stimuli. They conclude that the N170 may be part of a mechanism for face detection, or, given the enhancement when eyes are presented alone, a result of activation in an area sensitive to eyes.

Functional Sensitivity
The N170 is believed to reflect an early part of the face perception process. Some evidence for this is that familiarity of the individual in the photo has no effect on the N170, this was seen in studies by Bentin and Deouell, and by Eimer both in 2000. This means that the N170 is evoked at a point before recognition. It has therefore been put forward that the N170 reflects part of the process of the structural encoding of the face.

The 'face inversion effect', reported as early as 1969 by Yin, occurs when faces are presented upside down. Participants find recognition and perception of these faces more difficult than when they are upright. This effect is more pronounced in pictures of faces than in pictures of objects. The N170 is sensitive to face inversion, it has been found to be delayed by approximately 10 ms as compared to upright faces (Bentin et al., 1996 ). This result has been replicated in other studies for example Itier and colleagues published this result in 2006 when they presented faces, cars, houses, ape faces, eyes and chairs to participants. These stimuli were upright, inverted, distorted or inverted and distorted. They found that upright faces evoked an N170 that was delayed by about 10ms. The same delay was seen in the other non-face stimuli but not to the same extent. Itier et al. explain this delay as a changing of the rate of neuronal activation as a result of the unusual view of the face.

The amplitude of the N170 also increases when the face stimuli are inverted, this was seen in the Bentin 1996 study, where they reported a higher amplitude for inverted faces as compared to upright faces in the right hemisphere. This effect was replicated by Rossion and colleagues in 2000, they presented participants with images of upright and inverted faces, shoes, chairs, cars, greebles and houses. They found significant increases in amplitude in inverted compared to upright faces, this difference was not seen in any of the other stimuli. Itier and colleagues hypothesised that the reason for this increased amplitude is that when the faces are inverted there is an increase in eye-sensitive cell activation, this activation is thought to be inhibited when viewing upright faces. This hypothesis is based on the assumption that the N170 is generated by the superior temporal sulcus (STS), which is known to have neurons selective for eyes and faces. This hypothesis is supported by their research in 2007, in which they presented eyes, faces and faces without eyes. They found that when they inverted the faces without eyes the increase in amplitude normally associated with inverted faces disappeared. This hypothesis isn't without faults, as will be seen in the next section; the data on N170 localisation does not always show the STS to be the generator of the N170.

Localisation
There is some debate as to where exactly the N170 is being generated, the sides of the argument seem to be linked with two different methods of localisation: dipole fitting and computing distributed brain activation patterns. Although when considering this research it is important to keep in mind that source localisation in EEG is subject to debate as it can be unreliable. Bötzel and colleagues 1995 presented flowers, leaves and faces to participants. They used dipole fitting to try and discover the location of the negative waveform that responded selectively for faces. They found that their data indicated the occipito-temporal junction. Using the same method for locating the origin of the N170, Rossion and colleagues (2003) showed that their data indicated the inferior occipital gyrus with a right hemisphere bias, when they presented faces, words and cars to their participants. Deffke et al. (2007) use both EEG and MEG to attempt to locate the generator of the N170 and the M170. The M170 is the equivalent of the N170 but using MEG, they both occur at 170 ms and are thought to reflect the same neural processes. They only presented pictures of faces to the participants and their data indicted the fusiform gyrus as the origin of these waveforms. These three studies indicate an area that could be a part of the occipital face area (OFA), an area that has been linked to face perception in fMRI studies such as Gauthier and colleagues in 2000. Pitcher and colleagues used repetitive transcranial magnetic stimulation (rTMS) to look at face processing in the OFA. They applied rTMS over the right OFA, 6 double pulses 40 ms apart were delivered. Accuracy in a face part task only fell at the pulsed delivered at 60ms and 100ms. This indicated that the OFA is key to early face processing but does not support the idea that the OFA is the origin of the N170 as this occurs later.

The other method of localising the origins of waveforms from EEG data, by computing distributed brain activation patterns has led to conflicting results. Itier and Taylor conducted a study in 2004, in which they presented houses, mushrooms, tools, flowers, lions, road signs, textures and upright and inverted faces. Their data indicated that the source of the N170, when presented with faces, was the posterior STS, which supports their 2007 hypothesis regarding the increase N170 amplitude associated with presenting inverted faces (see the previous section). This conclusion is supported by MEG research, for example Watanabe and colleagues (2003) used both MEG and EEG to attempt to locate the origin of the M170/N170. They presented upright and inverted faces, scrambled images and object to the left and right hemifield participants and their data indicated that when presented with faces the waveform originates in the STS and the fusiform gyrus, with a bias towards the right hemisphere. Another MEG study by Henson and colleagues in 2007 confirmed this result, localising the MEG signal to posterior temporal areas and supported with EEG and fMRI data of their own from previous years showing the STS to be the origin.

N170 Abnormalities
Prosopagnosia is an inability to recognise faces, often brought on by brain injury but more recently developmental prosopagnosia has become recognised. There is a body of research looking at how the N170 differs from healthy participants in this group of people. Eimer and McCarthy looked at patient PHD in 1999, PHD became prosopagnosic as a result of a traumatic brain injury sustained in a car accident, he showed prosopagnosia and agnosia. They presented houses and unfamiliar faces to PHD and 24 healthy control participants. For the control participants, faces elicited a normal N170 over lateral temporal electrodes, but for PHD this waveform was missing. The authors go on to say that there were similar results when the participants and PHD were shown inverted faces and houses. They conclude that PHD's deficits are reflected in his ERPs to faces and that as PHD struggles with structural aspects of face processing that the N170 is a reflection of structural encoding. These results were supported by Kress and Daum who looked developmental prosopagnosic patients rather than patients with acquired prosopagnosia.

Bentin et al. 2007 look at this issue using both EEG and fMRI. They looked at patient KW, a developmental prosopagnosic, as well as 12 healthy controls who participated in the behavioural part of the study, 12 who participated in the fMRI part and 24 who were given EEG. In an identity task KW was slower and less accurate than controls. In terms of the fMRI data they found no areas in the fusiform gyrus or other ventrotemporal areas that were more sensitive to faces than images of places. But her parahippocampal place area was activating normally. The researchers reported that the healthy controls had normal face and place activation during this fMRI task. Finally KW's N170 did not differ between faces and watches unlike the N170's in the control group. This study further shows the link between an abnormal N170 component in prosopagnosics and their disorder but also supports it with abnormal activation patterns in response to faces as recorded by fMRI.

These studies show us that prosopagnosia is a deficit at the level of early face processing, this has also helped us to understand more specifically what process the N170 is a reflection of, that is structural encoding of faces.

N170 Face Specificity Debate
The most controversial study looking at the N170 in recent years has been that of Thierry and colleagues in 2007. They claimed that the higher amplitude seen in the N170 in response to faces as compared to other stimuli was not because the N170 reflects some part of face processing as was assumed by the many papers published on the subject but rather that it was an artifact caused by interstimulus variance. That is, that a particular face stimulus being presented to participants was more similar to other stimuli within that category than a particular non-face stimulus was to other non-face stimuli within their respective categories. They termed this concept 'interstimulus perceptual variance' (ISPV). They back up this claim in their first 2007 paper in which they report three experiments. In these experiments they manipulate the ISPV of the stimuli and look at the effect it has on the N170. In their experiments Thierry and colleagues presented participants with faces, cars and butterflies. They found that the N170 evoked by faces appeared earlier than cars but had a similar amplitude. They found a significant effect of ISPV, specifically that low ISPV had a higher amplitude that high ISPV. They claim that this shows that it is ISPV rather that face sensitivity that is causing the change in the N170 when comparing faces and objects. Their other two experiments were checking for effects of symmetry and attention on the N170 and found that they had no effect.

There were two main criticisms of this study, first that they have not correctly measured the N170 and second that past studies have controlled for ISPV. Bentin and a number of key researchers in the N170 field criticised the Thierry paper in 2007 in a letter to the editor of the journal in which it was published. They criticise Thierry for using inappropriate electrodes to measure the N170. They averaged the data from twenty electrodes, ten in each hemisphere. These were covering areas in which research has shown the N170 to diminish, such as more medial and superior areas, as well as medial occipital areas where Bentin et al. claim the N170 is sometimes absent. This averaging of inappropriate electrodes is, according to Bentin et al. the reason Thierry et al. report such small N170 amplitudes. In a reply to this letter Thierry published the data from the correct electrodes which showed no significant difference in amplitudes between cars and faces.

The second criticism that Bentin et al. make is that ISPV has been controlled for in previous studies. They go on to report the interstimulus pixel-wise correlations of the stimuli from the Rossion et al. 2000 paper that showed higher N170 amplitudes for faces than Greebles, cars and houses. The correlations showed that not only were all four categories highly correlated, houses being highest, showing that all the stimuli were very similar. But that these correlations were higher than the correlations Thierry and colleagues report which indicates a better control of ISPV.