Review evidence from fMRI and ERP studies for the specificity of brain systems responding to different emotional stimuli.
Functional Magnetic Resonance Imaging (fMRI)
fMRI is one of the most useful of tools to investigate the activities of the brain. An activation is the difference between distribution of signal intensities on two kinds of conditions.
Signal intensities differ according to the blood flow volume passing through a vessel and so they depend on the vessel diameter too (Fukami, 2007, p. 2460). Comparison of conditions at two points may be regulated using a magnetic resonance angiography (MRA) and correcting the signals by using the diameter. Hypothesis testing may use Rician distribution or Gaussian (Park, 2007, p.2472). The analysis of data allows us to locate brain activity while doing something like moving the finger or smelling or seeing. Recordings could be disturbed by movement of the head, cardiac and respiratory activity and noise from the scanner. Random noise also must be eliminated (Voultsidou, 2007, p. 99). The random matrix theory is suitable for data analysis
by some researchers (Voultsidou, 2007, p. 99).
The concept of mental rotation (MR) has been widely used in cognitive psychology and brain research (Christova, 2008, p. 79). This concept is used in functional neuroimaging studies like functional magnetic resonance imaging (fMRI) and positron emission tomography (PET) which document areas of the brain engaged in MR. Two features are commonly seen in all neuroimaging studies. They analyse the average of signals over the regions of interest (ROI). They also test whether a particular ROI showed a significant activation during an MR period as compared to a period of control (Christova, 2008, p. 79).
Most fMRI studies record the BOLD signal which is the blood oxygen level dependent signal using the GE EPI (gradient-echo echo-planar imaging ). BOLD sensitivity loss caused by susceptibility, induced in the readout direction, is featured (Weiskopf, 2006, p. 39).
A compensation approach is developed. The BOLD susceptibilty effects in the regions of neuronal activity causes microscopic magnetic alterations. GE EPI is sensitive to these alterations and to the macroscopic field in homogeneities which cause local image distortions and signal losses. The distortions can be corrected to a limited extent by online and offline methods. Signal dropouts are a problem compromising the fMRI studies in inferior frontal, medial temporal and inferior temporal lobes. The susceptibility related signal losses are reduced through several methods. The dropouts caused by field gradients in the direction of slice selection are focused upon (Christova, 2008, p. 79).Earlier fMRIs used thick slices while recently thinner slices are being used. Z-shimming is a widely used dropout compensation technique. It helps the formation of a composite image from a series of images with a different gradient pulse in the slice selection direction. These in-plane gradients in the phase encoding direction could also cause dropouts in the EPI by reducing the BOLD sensitivity. Different techniques are used to reduce dropouts. They are applying gradient pre-pulses in the Phase encoding (PE) direction, changing the slice orientation so as to reduce the PR component of the susceptibility induced gradient, reducing the EPI acquisition time and inverting the PE gradient polarity (Weiskopf, 2006, p.40). This can also be achieved by optimizing the slice tilt, PR gradient polarity and using the Z shim gradient prepulse. Even after optimizing, dropouts are seen with the orbito-frontal cortex and inferior temporal lobes. Strong gradients in the direction of readout lead to loss of signals in the EPI.
The asymmetry of the left-right and upper-lower visual fields can be studied in detail using the fMRI method and the BOLD response (Chen, 2004, p. 39). The reaction time is shorter and stronger when the lower visual field is stimulated than the upper. Similarly the right and left fields show different responses.
Audio-visual cross modal interactions
Perception of events in daily life rarely occurs through a single sensory modality (Belardinelli, 2004, p. 167). Processing of information from different senses leads to the perception. Due to the different modalities in an input, detection, localization and recognition or interpretation of the events become easier. Cross modal processing has given rise to two theories on neural pathways. One stresses the significance of multisensory cortical areas getting input from different senses. The second view is that synchronized firing causes a combined activity of the modality specific cortices. The heteromodal cortices in the human brain get the afferents from different senses as seen in animals (Belardinelli, 2004, p. 167). They are found in
the superior temporal sulcus, in the intraparietal sulcus, and in the prefrontal and limbic cortices.
They are found in the superior colliculus also below the level of the cortex. Research into the
regions found that the supra-additive response is enhanced by congruent audio-visual speech
when fMRI imaging was done. The left superior temporal sulcus showed strong interaction effects (Calvert et al, 2000). ). Another study showed that the facilitation effects were only for the visual targets on the auditory. Multisensory processing is also seen the right fronto-temporal area, anterior hippocampus and rostromedial orbitofrontal cortices (relationship between vision and smell) and right frontotemporal area (meaningless non verbal stimuli ). Another study compared the responses of the brain in crossmodal conditions which were more or showed a greater BOLD signal than the unimodal ones. The effect of the congruent cross modal presentations were mainly seen in the left parahippocampal gyrus and bilaterally in the visual unimodal cortices. It was inferred that the reactivation of the semantic cross modal associations was mediated by medial temporal structures (Belardinelli, 2004, p. 168). The left inferior frontal sulcus showed the effects of the incongruent crossmodal presentation.
Positron Emission Tomography and Applications in cancers
Management of various cancers requires the use of this potent imaging tool of PET
(Chatzifotiadis, 2006, p. 449). This is a nuclear medicine technique, originally used for quantitative imaging the metabolic activity of the brain. Intravenous positron emitting radio pharmaceuticals are injected. A PET scanner then images the flow of the medicine and its activity in the cancer regions. F-fluoride which has a short half live of 109 minutes the most commonly used positron emitter. It is used as F-fluoro-2-deoxy-d-glucose (FDG) to detect
the early steps of the accelerated glucose metabolism in cancers. The initial accumulation of the radiotracer into the cancer via membrane transport and its initial phosphorylation by hexokinase to FDG-6-phosphate are traced by the FDG (Chatzifotiadis, 2006, p. 449). However some cancers like prostate cancers, primary renal cancers, hepatomas, mucinous tumours and some low grade lymphomas do not have accelerated glucose metabolism and so PET would not be useful. Brain tumours have a high glucose acceleration but still cannot be imaged well due to the high uptake of FDG by the background. The tissues which uptake glucose are the brain, heart,
kidneys, testes, exercising skeletal muscle, and the kidneys. Excretion is via the kidneys so renal lesions are not easily imaged. PET is a molecular imaging tool and provides quantitative data based on the amount of radioactivity in the tissues and is non-invasive in nature (Chatzifotiadis, 2006, p. 449). Lesions less than one cm. but more than 5mm. may be detected. The detection depends on the degree of uptake against the clarity of the background. Magnification of images is not possible. If the background is also uptaking the tracer, a hot spot or a place of malignancy may be detected but the size may not be clear. PET is being combined with CT or MRI as single machines through software fusion to detect the exact site and size of the lesion. PET has better imaging than either CT or MRI. Tumour staging, assessment of response to treatment and restaging in recurrence are the most common uses (Chatzifotiadis, 2006, p. 450).
EEG is spontaneous electrical activity at the scalp. The signal is the electrophysiological response. Electrical dipoles in the pyramidal cell layer produce the electrical potentials recorded as EEG. Localisation of these dipoles is the main problem in the study of brain activity with the EEG. Human brains have the possibility of being connected to “to the intelligent computing applications in form of brain computer/machine interfacing (BCI/BMI) technologies” (Rutkowski, 2008, p. 122). One form of brain electrical activity is the Electro encephalogram (EEG). This can be captured and used for the BCI/BMI applications. However noise disturbances caused by electrophysiological signals and from the different machines in the environment could produce electromagnetic disturbances. Computer aided communication has made the non-invasive EEG more acceptable. Steady state potentials produce steady responses which can better evaluated have increased the popularity of EEG. EEG allows the “detection, estimation, interpretation and modeling of brain activities, and cross-user transparency” in the signal processing (Rutkowski, 2008, p. 122). This technique is now seen at the centre of future “intelligent computing”. Industries which could benefit from the online analysis and visualization of the brain states are the prosthetics, entertainment and the computer games. EEG is a summation of the post synaptic potentials from a number of neurons in the brain. They can be monitored and classified after giving a stimulus to the patient. These are divided into components called intrinsic mode functions (Rutkowski, 2008, p. 123).
Electrical artifacts can be caused by eye movement in the process of taking an EEG.
Methods like regression in the time domain or frequency domain have been adopted to prevent these artifacts (He, 2007, p.495). Newer techniques include the principal component analysis
(PCA) and independent component analysis (ICA). A recent method involves an adaptive filtering technique. This method appears to be useful in removing the horizontal and vertical components of the artifacts. A noise canceller can be also added (He, 2007, p. 495).
Epilepsy is evaluated using the EEG. The spikes in the EEG are diagnostic of the illness.
Idiopathic focal epilepsy sees a difference in the seizures and EEG changes at puberty. However psychomotor disturbances remain and renders the patient severely affected (Gross-Selbeck, 2004, p. 90). Considerable impairment of speech and mental development and a regression on the illness could be seen in the EEG. The question remains as to whether to continue treating a person who had abnormal paroxysms in the EEG but whose EEG shows a change for the better. It is not proven that this person with a residual psychomotor disturbance can go on without treatment . The conclusion in this study was that a person with idiopathic focal epilepsy need not have treatment other than for his seizures to improve neuropsychological outcome (Gross-Selbeck, 2004, p. 90). Treatment should be given to children with epilepsy of less than 6 years duration, extensive changes in the EEG and IQ less than 60. The EEG should be used to decide about treatment at frequent intervals. A child with more than six years of illness requires treatment only if there a regression of the mental status (Gross-Selbeck, 2004, p. 90). Partial epilepsy due to neuronal migration disorders were compared to patients with brain tumours (BT) and hipocampalsclerosis (HS) (Degen, 2004, p. 22). The clinical, MRI findings and EEG findings were compared in these surgically treated patients. The MRI showed temporal localizations of the brain abnormalities in BT and HS patients. The interictal EEG was 95% and the ictal EEG was 85% in BT and HS patients. Extratemporal localization was found in NMD by 58%. NMD patients showed more “retarded developmental milestones (46.3%), intellectual deficits (78.4%), behavioral disorders (68.3%), neurological deficits (61%) and a more unfavorable surgical outcome (class I A-D:37.5%)”. (Degen, 2004, p. 22). More favourable findings were seen in BT. Partial seizures were therefore concluded to be of the less favourable type.
Magneto encephalography (MEG)
Technology has allowed scientists to look inside a living brain (Kim, 2006, p. 2714).
MEG is a non-invasive technique which “measures the magnetic field induced by neural current flows in the cerebral cortex result of electrical activity in neural cell assemblies”. The neural current sources from a measured distribution of an electromagnetic field are measured by the MEG source localization. The equivalent current dipole method is used. The advantages are that it is easy to implement and robust to various noises (Kim, 2006, p. 2714). A variety of techniques are available for MEG imaging data to the corresponding cortical structures. MR-FOCUSS (Multiresolution Focal underdetermined system solution) allows spatial resolution to be “selected appropriately for focal or extended sources”( Moran, 2005, p. 1). The cortical activity is enhanced to get the final high resolution results as in the FOCUSS imaging technique. Quantification is possible. Language processing for pre-surgical planning is possible and effective with MR-FOCUSS. Sequential activation of many co-related sources in the language processing is also possible ”( Moran, 2005, p. 1).
Belardinelli, M.O. et al. (2004). “Audio-visual crossmodal interactions in environmental perception: an fMRI investigation” Cogn Process (2004) 5: 167–174
DOI 10.1007/s10339-004-0024-0, Springer-Verlag 2004
Chatzifotiadis, D.N. et al, (2006). “Positron Emission Tomography and Cancer”. in Oncology : An evidence based approach. (Eds.) Alfred Chang et al, 2006 Springer Science+Business Media, Inc.
Chen, H. et al. (2004). “A Study on Asymmetry of Spatial Visual Field by Analysis of the fMRI BOLD Response”. Brain Topography, Volume 17, Number 1, Fall 2004
Christova, P.S. et al (2008). “A voxel-by-voxel parametric fMRI study of motor mental
rotation: hemispheric specialization and gender differences in neural processing efficiency Experimental Brain Research, Vol. 189. DOI 10.1007/s00221-008-1405-x
May 2008, Springer-Verlag 2008
Degen, R. (2004). “Clinical, MRI and EEG findings in surgically treated patients with partial epilepsies due to neuronal migration disorders (NMD) in comparison with those of patients with brain tumors (BTU) and hippocampal sclerosis (HS)”. Z Epileptol 17:22–29 (2004) DOI 10.1007/s10309-004-0043-y
Fukami, T et al. (2007). “Correction of fMRI signals by using the vessel diameter extracted
from MRA images” in World Congress on Medical Physics and Biomedical
Engineering 2006, (Eds.) Sun I. Kim and Tae Suk Suh (Eds.), Springer Berlin Heidelberg New York
Gross-Selbeck, G. (2004). “EEG in idiopathic focal epilepsy– when to treat?”
Z Epileptol 17:90–102 (2004) DOI 10.1007/s10309-004-0050-z
He, P. (2007). “Removal of ocular artifacts from the EEG: a comparison between
time-domain regression method and adaptive filtering method using simulated data”.
Med Bio Eng Comput (2007) 45:495–503 DOI 10.1007/s11517-007-0179-9 International
Federation for Medical and Biological Engineering 2007
Husmann, L. et al, (2008). “Diagnostic accuracy of myocardial perfusion imaging with
single photon emission computed tomography and positron emission tomography: a comparison with coronary angiography”. Int J Cardiovasc Imaging (2008) 24:511–518
DOI 10.1007/s10554-007-9288-7 Springer Science+Business Media B.V. 2007
Kim, B.J. (2006). “Enhancement of calculation time and accuracy in MEG inverse problem using simplex method” in World Congress on Medical Physics and Biomedical
Engineering 2006, International Federation for Medical and Biological Engineering 2007
Moran, J.E.et al, (2005). “Multiple Resolution FOCUSS : A source imaging technique applied to MEG data”. Brain Topography, Vol. 18, Springer Science and Business Media
Park, C. et al. (2007). “Use of the Rician Distribution with the EM Algorithm for Hypothesis
Testing of fMRI Data”. in World Congress on Medical Physics and Biomedical
Engineering 2006, (Eds.) Sun I. Kim and Tae Suk Suh (Eds.), Springer Berlin Heidelberg New York
Rutkowski, T.M. et al. (2008). “EMD Approach to Multichannel EEG Data – The Amplitude and Phase Synchrony Analysis Technique”. D.-S. Huang et al. (Eds.): ICIC 2008, LNCS 5226, pp. 122–129, 2008. Springer-Verlag Berlin Heidelberg 2008
Voultsidou, M. et al. (2007). “Feature evaluation in fMRI data using random matrix theory”.
Comput Visual Sci (2007) 10:99–105 DOI 10.1007/s00791-006-0037-6
Weiskopf, N. et al, 92006). “Optimized EPI for fMRI studies of the orbitofrontal cortex:
compensation of susceptibility-induced gradients in the readout direction”.
Magn Reson Mater Phy (2007) 20:39–49 DOI 10.1007/s10334-006-0067-6