Copy to URL Repository Edition
- DTI-based fibre tracking in the examination of visuomotor networks (Jan Scholz)
- Functional connectivity in healthy subjects with auditory verbal hallucinations (Kelly Diederen)
- Effects of sentence context in L2 natural speech comprehension (Ian FitzPatrick)
- The faster one has blinkers on: The role of co-representation in response inhibition and error detection (Stephan Miedl)
- Language lateralization in healthy people with Auditory Verbal Hallucinations: A pilot study (Kelly Diederen)
- Auditory Selective Attention as a method for a Brain Computer Interface (Michiel Kallenberg)
- The effect of fame as a context - An fMRI study (Gitty E. Smit)
- Localisation Techniques to improve Analysis for BCI (Rianne Hupse)
- What the electrocorticogram can tell us about the electroencephalogram (Denise van Barneveld)
- Contextual memory in patients remitted from their first episode of depression: An fMRI study (Sara Pieters)
- Neural Correlates of Masking (Martijn Schippers)
- The Interplay of Prosody and Syntax in Sentence Processing: Two ERP-studies (Sara Bogels)
DTI-based fibre tracking in the examination of visuomotor networks
The ability to follow axonal tracts in white matter using diffusion tensor imaging (DTI) has existed for some years. Only very recently have research groups attempted to combine functional magnetic resonance data (fMRI) with tractographic data obtained with DTI. In this project we assessed the utility of DTI-based fibre-tracking for examining visuomotor networks in healthy human adults. We hypothesized that cortical pathways for oculomotor control can be mapped using a combined DTI-fMRI approach. We used a delayed-saccade task to map the topographic organization of the frontoparietal oculomotor system. Varying the polar angle of the peripheral target for a delayed saccade in a systematic fashion revealed several topographically organized areas in parietal and frontal cortices. Subsequently we attempted to determine anatomical connectivity of these regions, in particular the connections between the human analogues of monkey areas FEF and LIP, with several DTI-based fibre-tracking approaches. Each approach met with limited success.
Auditory verbal hallucinations (AVH) are a core feature of schizophrenia. Previous studies have provided evidence for dysfunctional integration of language regions to underlie AVH. However, schizophrenia is a complex syndrome consisting of psychotic, cognitive and negative symptoms. In order to learn whether this mechanism plays a causal role in the pathophysiology of AVH, the “pure” form of AVH should be investigated, which can be found in healthy subjects with AVH. Functional integration was assessed with psychophysiological interactions (PPI’s) in 10 healthy subjects with AVH, 10 schizophrenia patients and 10 healthy controls matched for age, handedness and education. Subjects were scanned while covertly performing a paced letter fluency task. Schizophrenia patients and healthy subjects with AVH displayed dysfunctional connectivity between the left anterior cingulate cortex (ACC) and the left dorsolateral prefrontal cortex (DLPFC) and between the left inferior frontal gyrus (IFG) and the left superior temporal gyrus (STG). Additionally, dysfunctional connectivity between the left ACC and the left STG was found in the schizophrenia patients. These findings suggest a dysfunction between the production and the perception of speech in healthy subjects with AVH and schizophrenia patients. Furthermore, the difference between healthy subjects with AVH and schizophrenia patients provides an explanation why healthy subjects with AVH can execute some control over their hallucinations, in contrast to schizophrenia patients.
Keywords: Auditory verbal hallucinations, language, functional connectivity, fMRI
In this study we investigated the effect of sentence context and both intra- and interlingual initial phonemic overlap on word recognition in L2 natural speech comprehension, using Event-Related Potentials (ERPs). Dutch(L1)-English(L2) bilinguals listened to English sentences in which the sentence-final word was: (1) semantically fitting, (2) semantically incongruent, (3) semantically incongruent, but sharing initial phonemes with the most probable sentence completion within the L2 (thus initially congruent with the sentence context), (4) semantically incongruent, but sharing initial phonemes with the L1 translation equivalent of the most probable sentence completion. We found an N400 effect comparing each of the semantically incongruent conditions to the semantically fitting condition, demonstrating a clear effect of semantic incongruity in non-native natural speech comprehension. The peak latency of the N400 component seems to be later in non-native listeners compared to monolinguals. In contrast to previous findings (Connolly & Phillips, 1994; Diaz & Swaab, in press; Van den Brink, Brown, & Hagoort, 2001, 2006; Van den Brink & Hagoort, 2004; Van Petten, Coulson, Rubin, Plante, & Parks, 1999) we found no evidence of an early negativity comparing the semantically incongruent and initially congruent conditions. We did find a significant delayed peak latency of the N400 component in the within language overlap condition. This finding demonstrates that semantic integration in non-native listening can start on the basis of the word-initial phonemes. We found no evidence of a delayed N400 in the between language overlap condition.
Earlier experiments have shown that the formation of a co-representation of the task of others influences ones own action behaviour. The aim of the present study was to investigate how differences in co-representation, while doing a go-nogo task together with a second participant, influence ones own action-monitoring processes. The results showed a smaller NoGo P3, more errors and a reduced ERN on stimuli requiring only the response of the other participant compared to stimuli requiring the inhibition of both participants. Moreover, significant interactions with competitive response strategy revealed that the monitoring processes of fast responders were least affected by shared action representations. These results create evidence that the existence of a co-representation in a joint task leads to specific modulations of action-monitoring processes. Interestingly, these effects may also depend on the response strategies people employ in a competitive setting.
Language lateralization in schizophrenia has been reported to be decreased, which is due to increased language activity of the right hemisphere. Furthermore increased right hemisphere language activity is correlated to the severity of auditory verbal hallucinations (AVH) in schizophrenia, and could possibly be a predisposing factor for AVH. However, schizophrenia is a complex syndrome consisting of psychotic, cognitive and negative symptoms. In order to learn whether decreased lateralization plays a causal role in the pathophysiology of AVH, the isolated form of AVH should be investigated, which can be found in healthy subjects with AVH. Language lateralization was measured in 10 healthy subjects with AVH and compared to 10 schizophrenia patients and 10 healthy controls matched for age, handedness and education. Subjects were scanned while covertly performing a paced letter fluency task. A significant main effect for group was revealed with respect to the right hemisphere. This was due to increased activation in the right hemisphere of the schizophrenia patients relative to the healthy controls. Differences in language lateralization between schizophrenia patients, healthy subjects with AVH and healthy controls did not reach significance. However, mean lateralization indices were lower in both hallucinating groups. This difference is likely to become significant when sample size increases
Keywords: Auditory verbal hallucinations, language lateralization, fMRI.Auditory Selective Attention as a method for a Brain Computer Interface
The object of this study was to design a Brain Computer Interface (BCI) based on auditory selective attention (ASA). ASA is a promising paradigm for a BCI, as focusing attention does not requite a lot of training, whereas the possibility of offering a large number of possible targets facilitates a high bit rate.
In this study subjects focused attention on one tone out of two. The two tones were separated in space and pitch, and each tone was frequency tagged by means of amplitude modulation (AM). AM tones are known to evoke an auditory steady state response (ASSR) at the am frequency, and previous research has demonstrated that the power of this ASSR is increased by selective attention.
To detect the direction of the subject’s attention, features were calculated that characterized the ASSR. Subsequently, a classifier was trained based on linear discriminant analysis.
The best results were obtained with a feature that consisted of the real and imaginary parts of the fourier transformed signal at the am-frequency. Electrodes above the auditory cortices yielded the best results. On perception data, single trial classification reached a classification rate of 80%. On attention data, the best classification rate was 68%.
The current BCI achieved a bit rate of 3.78 bits/min, which is moderate compared to other BCI-systems. We will discuss several procedures for improvement.
Keywords: Brain-Computer Interface, auditory selective attention, frequency taggingThe effect of fame as a context - An fMRI study
Gitty E. Smit
Over the past decades there has been an increase of celebrities in advertising, but there is still a huge variance in the effectiveness of celebrities in advertising. Previous studies (Klucharev et al., 2006; Rossiter and Smidts, 2006) have shown that memory and attitudes of a product increases when the famous presenter is an expert on the presented product. To study the mechanisms of effective use of famous presenters in advertising we simulated advertising and studied the modulation of memory and attitudes for products with fame as a context. We presented 24 female subjects with photos of products (shoes) coupled to famous and non-famous faces. To contrast effects of presenters’ attractiveness and expertise as well as specific item characteristics, we equated attractiveness and used only shoes as stimuli. During this task we recorded brain activity via fMRI. We found a substantial behavioural effect of fame on memory for the products; more products coupled to famous faces were remembered. Turning to the neural underpinnings of this behavioural effect, we found a subsequent memory effect x fame interaction in the thalamus. Moreover, we found a main effect of fame in left frontal and temporal regions, which are commonly seen during semantic processing and a positive emotional encoding context. Celebrities trigger semantic knowledge and can be seen as a positive emotional encoding context. Behaviourally they increase memory for the products presented. It seems that using a celebrity in advertising increases brand awareness and memory of the advertised product by increasing elaboration and positive emotional processing of the advertised product such that brand awareness increases.
Key words: memory encoding, subsequent memory effect, attitude change, subsequent attitude effect, encoding context, fame, celebrities, faces, neuroeconomics.Localisation Techniques to improve Analysis for BCI
An experiment is described to explore the possibility to elaborate a new type of brain computer interface (BCI) system: a system in which selective attention to rhythmic tactile stimuli is used to control an external device. It is known that temporal rhythmic tactile stimuli induce a steady-state somatosensory evoked potential oscillating at the same temporal frequency as the driving stimulus. The amplitude of this oscillation increases when the subject is attending to the stimulus. This attention induced power gain can be detected in the EEG (electroencephalogram) and might be translated into commands for a computer or other device. Because EEG data has a small signal to noise ratio, it is investigated if a beamformer spatial filter improves the classification success rates. To be sure that the beamformer filter allocates activity to the correct anatomical locations, a new method is proposed to construct realistic head models. During the experiment, tactile stimuli with different temporal frequencies are presented to the left and right index finger of a subject. The subject is instructed to attend to one finger and to ignore the sensations of the other finger. Perception conditions, in which only one finger was stimulated, were included as a baseline. Single trial EEG and voxel data was used in a classification algorithm to detect which finger was attended. Classification rates are relatively high (± 90%) for perception conditions, while selective attention conditions give success rates below chance level. An explanation for this can be that the attention induced power gain is too small to be detected in single trials. Results show that beamforming allocates the frequencies presented to left and right index finger to separate areas of the brain. This is in contrast to electrode data, in which the two frequencies reach electrodes at both sides of the scalp. Therefore, by using the beamformer filter the power of a single stimulus is focused to a single area which will increase the signal to noise ratio. However, classification success rates show no improvements when the beamformer filter is used. A suggestion for further research is to include a discrimination task in the experiment which might increase the attention level of the subject and therefore the attention induced power gain. The amount of trials can be increased that is used as a training set in the cross validation procedure of the classification algorithm. Furthermore, the amount of trials used for estimating the covariance between the sensors, which is necessary for building the beamformer filter, might improve the results. More experiments are necessary to investigate if the results described here are representative to sessions with other subjects.
Denise van Barneveld
Although brain signals measured under the skull (electrocorticogram, ECoG) and signals measured on top of the scalp (electroencephalogram, EEG) stem from the same brain activity, they are different. We investigated how we can produce EEG when we know ECoG (“forward problem”) and how we can produce ECoG when we know EEG (“inverse problem”). We modeled the head as three concentric spheres, representing the brain, skull and scalp. Brain activity is simulated by a dipole. The forward method links the ECoG potentials on the inner sphere to the EEG potentials on the outer sphere via a transfer matrix, based on the geometries and the conductivities of tissues involved. Results showed that the error between analytically computed EEG and EEG produced from analytically produced ECoG with the forward method, is smaller at electrodes close to the source, compared to electrodes far away from the source. The higher the resolution of an ECoG electrode grid, the better the forward model works. Another finding was that the forward model is more accurate or surface sources, compared to deep sources. This result is of practical importance, since most cognitive interesting sources stem from the cortex (the outermost layer of the brain). In the inverse model, the transfer matrix is inverted and additional regularization constraints are applied to compute ECoG from simulated EEG. We showed that the inverse model gives good results. The forward method is tested with data measured from an epileptic patient at the University of Freiburg. Results show that the forward model gives better results at the EEG electrode overlying the ECoG grid compared to the electrode posterior to the grid. Further research is needed to make errors smaller.
Keywords: Electroencephalogram, Electrocorticogram, Three-sphere model, Forward problem, Inverse problem, Brain Computer InterfaceContextual memory in patients remitted from their first episode of depression: An fMRI study
Background: Memory problems are a well known symptom of major depressive disorder (MDD). Recollection of past experiences is thought to be more impaired than familiarity in MDD. We investigated whether these memory problems are still persistent in remission. Furthermore, we tried to delineate the neural correlates of memory processing in remission.
Methods: 13 remitted patients (non medicated, remitted from first episode) and 13 matched controls participated in a contextual memory task while lying in a fMRI scanner.
Results: Behavioural performance did not differ between remitted patients and controls. We did find a difference between remitted patients and controls in brain activation during encoding.
Conclusions: Although the sample size is small, there is striking evidence that MDD patients even when being in remission from their first episode already form declarative memories differently.
This study looks at the phenomenon of visual masking. When two stimuli are presented in rapid succession the perception of one can be blocked by the perception of the other, depending on the timing and features of the stimuli used. There is currently an open discussion as to whether this effect is neurally limited to brain activity in the primary visual areas or whether it also extends to higher parts of the brain. We try to determine the neural correlates of visual masking by showing where in the brain there is a significant difference between consciously (unmasked) and not consciously (masked) perceived stimuli.
Two behavioral experiments were performed to determine the optimal stimulus-setup and stimulus-timing for masking effects. The optimal settings found were used in an fMRI experiment. The stimulus visibility was measured while manipulating the time between the start of the stimulus and the start of the mask. This Stimulus Onset Asynchrony (SOA) influences the visibility of the stimulus in a manner that looks like a U-shape (high visibility at short and long SOAs, low visibility in between). We added a control condition where we replaced the mask with a non-masking, but physically almost identical shape. As this control shape does not mask, the SOA should not influence visibility, and hence a U-shape should not occur when plotting visibility versus SOA. These two differently shaped masking functions make it possible to attribute differences in brain activity (as measured by fMRI) to a difference in visibility.
As our behavioural data in the scanner did not lead to such opposing masking functions, the found brain activity differences could not be attributed to an effect of masking. We did find activity in the primary visual cortex, the left angular gyrus and extrastriate cortex and the left superior temporal gyrus, but this data can only be taken as an indication where activity might be found in a study where the behavioural data is significantly different. As we present this study as a pretest for such a more elaborately controlled experiment, we present some suggestions as to how this level of control over the masking effect can be achieved.
In earlier studies, which were mostly reading studies, it has become clear that not only syntax but also other factors such as semantics and discourse context play an important role in sentence processing. Much less research has been done to investigate auditory sentence comprehension, although this is by far the most common way of human communication. The focus of the present study, prosody, is unique to auditory sentence processing. ERPs are presumably the best method to investigate auditory sentence comprehension, because they are the only straightforward, online method available. In this thesis, two experiments are described that used two different types of locally ambiguous sentences to investigate the role of prosody in sentence processing.
In Experiment 1, as a follow-up on Kerkhofs et al. (submitted-a), sentences with an NP/S-coordination ambiguity with and without a prosodic break (PB) were used (see sentences 1 and 2).
1. The mannequin kissed the designer (PB) and the photographer on the party. (NP)
2. The mannequin kissed the designer (PB) and the photographer opened a bottle of Champaign. (S)
According to late closure, the NP-coordination sentence should be preferred. However, we hypothesized that the PB could reverse this preference. At the PB, a Closure Positive Shift (CPS) was found, replicating Kerkhofs et al. Comparing the S-coordination sentences with and without a PB at the disambiguation point (opened), a mid-frontal P600 effect, which indicated processing difficulty, was found throughout the experiment. This result was different from Kerkhofs et al., who found a LAN-effect in the first half of their experiment. This difference could have been caused by a different ratio of items in the different conditions in the two studies. Comparing the NP-coordination sentences with and without a PB at the disambiguation point (on the party), a comparison that was not included in Kerkhofs et al., a mid-posterior P600 effect was found for the sentences with a PB, but only in the first half of the experiment. This asymmetrical pattern of effects is difficult to interpret. One possible explanation is that participants came to regard sentences with a PB in a special way in the course of the experiment.
In Experiment 2, as a follow-up on Steinhauer et al. (1999), sentences with another type of late closure ambiguity with and without a PB were used (see sentences 3 and 4).
3. De verpleegster hielp (PB) de zieke te lopen… The nurse helped the patient to walk…
4. De verpleegster hielp (PB) de zieke te vervoeren… The nurse helped to transport the patient…
According to late closure, sentence 3 should be preferred because de zieke (the patient) is the object of the previous verb hielp (helped). However, we hypothesized that a PB after hielp could reverse this preference. At the PB, the ERPs showed a CPS, which replicates Steinhauer et al. Comparing the sentences with a PB at the disambiguation point, an N400 effect was found for the condition with a mismatch between prosody and syntax (3). This contrasts with Steinhauer et al. who found a biphasic N400/P600 response in a similar comparison in German. The fact that we only find an N400 effect (without a P600) suggests that here, the PB is such a strong cue for a certain syntactic parse, that - in case of a mismatch - the disambiguating verb is picked up as a semantic anomaly, without triggering a revision of the incorrect syntactic analysis. Comparing the sentences without a PB at the disambiguation point, a LAN-like effect for the mismatch condition (4) was found, but only in the first half of the experiment. This suggests that listeners can use the absence of a PB in a strategic way while this appears to be impossible for the presence of a PB.
Overall the results suggest that the CPS is a reliable indicator of a PB. At the disambiguation point, the results of the two experiments are quite different. Both experiments show that prosody can influence the decision to analyze the sentence in a certain way, at least initially. Furthermore, the differences between the results of the two experiments make clear that the nature of processing difficulty in late closure sentences depends on the precise nature of the structures themselves.