Introduction and Background
Summary
The purpose of this research project is to use Magnetoencephalography (MEG) to non-invasively localize brain regions that facilitate the control of attention and other higher executive processes. In general, these data will be of clinical use in determining if there are cortical activation differences in subjects with Attention-Deficit Hyperactivity Disorder (ADHD). The location and sequence of activation of discrete cortical regions involved in attention and the executive control of behavior will be imaged. Cortical activation associated with attention and executive control will be stimulated using well-established paradigms (Continuous Performance Test (CPT) and Stroop Interference Test). MEG imaging studies will also be performed on control subjects to compare with the results obtained from ADHD subjects. This comparison will determine if the localized areas of control are the same and, if not, how they differ. MEG data analysis will be performed utilizing single equivalent current dipole software (Magnetic Source Imaging) and current density imaging software (Multi Resolution FOCUSS). These analyses will determine the latency, amplitude, and sequence of activation associated with each subject’s brain structure utilizing their MRI scans. The main aim of this study is to establish the efficacy of specific MEG imaging techniques in determining the structure, activation sequence, and strength of neuronal interaction during attention and executive control. This study of MEG neuroimaging will increase the understanding of how attention and other forms of executive control occur in those diagnosed with ADHD. In addition, the basic science of validating neuronal network models of attention and executive control in behavior will be extended. Finally, the results of this study may help refine the diagnosis of subtypes of ADHD, leading to selective and more effective behavioral and pharmacological treatment of these subtypes.
Specific Aim
To use Magnetoencephalography to image the location of cortical processes associated with attention and executive control over behavior in subjects with Attention-Deficit Hyperactivity Disorder and in Normal Control subjects, and also to determine the latency and sequence of activation of these cognitive neural pathways.
Hypotheses
The major hypotheses for the proposed research were derived from a range of brain imaging studies focused on ADHD as well as the co-investigator’s previous research on the neuropsychology of attention and arousal [1-12]. These hypotheses are provided in the table below and organized according to the flow of processing that underlies attention and other executive functions. The table provides the neural region hypothesized to be involved in each stage of processing, the experimental measure to be used for each stage, and the results predicted for the ADHD subjects.
3. Proposed Study
A whole head Neuromagnetometer will be used to measure magnetic fields in 5 subjects with ADHD and 5 age-matched control subjects (no neurological disorders) between 18 and 24 years of age. After each subject has signed an informed consent, measurements will be taken inside a magnetically-shielded room located in the Neuromagnetism Laboratory at Henry Ford Hospital (HFH). Neuromagnetic recordings of Visual Evoked Cortical Magnetic Fields (VECMFs) [13] will be acquired to determine the landmarks of the brain for each individual. A Visual Continuous Performance Test (CPT) and the Stroop Interference Test will be used to induce the attentional and executive control processing that is of interest in this study. Prior to MEG testing each subject will be interviewed and complete the questionnaires as discussed below.
Subject preparation for MEG study:
Each subject will change into a hospital gown and remove all metal articles from his/her body. Three small electrode coils, used to transmit subject location information to the neuromagnetometer probe, will be taped to the forehead with two-sided tape. Disposable ear molds of the correct size will be placed in the ears and an additional localization coil will be attached to each ear mold. A commercial videotape eraser will be used to demagnetize dental work. The subject will then lie comfortably on the bed, and automatic probe position routines will be used to locate the head with respect to the neuromagnetometer detector coils. The neuromagnetometer helmet will be placed around the subject’s head so that the detector array will be in close proximity to most of the cortical surface. The subject will be asked to remain in a fixed position for intervals of 4-6 minutes during monitoring. With time to place a subject in the room and ensure his/her comfort, the complete testing will last approximately two hours, followed by one hour for the MRI scan to be performed.
The conventional method for source localization in clinical MEG studies has been to model the magnetic field as a single equivalent current dipole (ECD) [14,15]. This methodology works well for stationary, non-distributed sources such as early cortical latencies in evoked response data. For spontaneous transients, such as the cognitive functions of interest, the model will not be robust. In part this is due to the fact that there are significantly different areas of neuronal activity involved in a short period of time producing non-stationary distributed sources. Under these circumstances, other modeling techniques are more useful. Multi-Resolution FOCUSS (MR-FOCUSS) [16] is a current density imaging technique that has been developed at HFH and shown to be useful for imaging focal and extended sources. We propose to apply this technique to MEG data during the Continuous Performance Test and the Stroop Interference Test.
1. Questionnaires & Interview Clinical Diagnostic and Personality Measures
The Structured Clinical Interview for DSM-IV-TR (SCID [17]), a Neurological and Medical Screen, and the Mini-Mental State Exam [18] will be given to establish the psychiatric, medical, and cognitive status of prospective subjects for the purpose of establishing inclusion and exclusion criteria.
The Wender-Utah Rating Scale is a self-rating instrument for adults of childhood behaviors. This scale aids in the retrospective diagnosis of childhood ADHD [19,20]. It will be used in the present study to exclude a person with childhood ADHD symptoms from the normal control group as well as to establish the adult ADHD diagnosis and possible ADHD subgroups.
The Conner’s Adult ADHD Rating Scale [21] will be used with the Wender-Utah Rating Scale to help establish the diagnosis of ADHD for the self-referred and physician referred volunteer subjects to be tested in this study.
The Speilberger State-Trait Anxiety Scale (STAI) [22] is designed to differentiate between the temporary condition of "state anxiety" and the more general and long-standing quality of "trait anxiety" in adolescence through adulthood. The State Anxiety scale evaluates feelings of apprehension, tension, nervousness, and worry, which increase in response to physical danger and psychological stress. Anxiety levels are known to affect both the CPT and SCR dependent variables and it is, therefore, necessary to account for level of anxiety in the analysis of these results [18].
The Tridimensional Personality Questionnaire (TPQ) is a 100 item, self-administered paper and pencil, true-false instrument that measures three higher-order personality dimensions, which influence how a person responds to stimuli. These dimensions have been linked to the brain systems and neurochemistry of importance to this study [24-28]. The analysis of the genome of these personality traits is currently underway [28]. The three personality dimensions measured by the TPQ include: 1) Novelty Seeking or behavioral activation as maintained by dopamine, 2) Harm Avoidance or behavioral inhibition as maintained by serotonin, and 3) Behavioral Maintenance or reward dependence as maintained by norepinephrine [25-28]. This test will take approximately 10 minutes to complete. This questionnaire will be used when sufficient data exist to form behavioral/personality subgroups within the ADHD diagnosed subjects. These TPQ-distinguished, ADHD subgroups will be used to test hypotheses that predict different CPT and Stroop profiles based on the different underlying neurochemical profiles.
2.Visual Evoked Cortical Magnetic Fields (VECMFs) will be measured during the presentation of an alternating (1Hz) square-checker board pattern for 2 minutes [13]. This pattern will be viewed on a projection screen. Each subject will be asked to lay still, keep eyes open and focus on a red dot in the center of the image on the screen. The right half of the visual field will be stimulated for 2 minutes, and then the left visual field will be stimulated for two minutes. Both trials will be repeated. VECMFs will be band pass filtered (3-70 Hz), digitized (250Hz), and 200 Epochs will be averaged. Each epoch contains one cycle of a black and white alternating checkerboard pattern flash. The largest responses will be selected and ECD dipole fits made to localize the cortical locations where these sources arise. These peaks will occur at the time ~75 ms, ~100 ms, and ~145 ms after visual triggers. VECMF will be used to determine the cortical location of visual processing.
3. A Visual Continuous Performance Test [29] will be used to measure the subject’s MEG field responses while he/she maintains vigilance. Random letters will be shown for a brief 150 ms each with a 1.8 second inter-stimulus interval. The subject will be instructed to respond as quickly possible by pressing a switch to the “X” target stimulus when it follows the “A” cue stimulus. Four blocks of 100 trials each will take approximately 12 minutes to administer. Twenty percent of the trials in each block will be cued, “no go” trials (the “A” cue appears, but is not followed by the target “X”) and 20% of the trials in each block will be target, “go” trials (the “A” cue is followed by the “X” target. Cortical activation between 0-650 ms will be analyzed by ECD and MR-FOCUSS, to determine the latency and source of neuronal activity that accompanies the various experimental trial conditions of interest. These include (but are not restricted to) the neuronal activity associated with errors of omission (no response made for a target “go” trial) and errors of commission (response made for a cued, no target, “no go” trial),
4. The Stroop Interference Test [30] will be used to measure the subject’s MEG field responses during a task that necessitates the use of higher executive control over behavior. There will be 3 conditions in this test. Each condition will consist of 40 trials. Stimuli will be presented individually for 500 ms, with a 1.5 second inter-stimulus interval. The first condition involves the presentation of small rectangular blocks of color. The subject will be asked to quickly verbalize the name of the color presented. The second condition involves the presentation of the names of colors in black text. The subject will be asked to read the word presented as quickly as possible. The final condition is the interference test, in which the words are presented in a text color that is incongruent to the name of the color. The subject will be asked to name the color of the text as quickly as possible, not to read the word. Interference in processing occurs because we automatically default to the reading of the word. An accurate identification of the color involves the active inhibition of this over learned response. The first two conditions serve as control comparisons for the interference condition. Cortical activation between 0-650 ms will be analyzed by ECD and MR-FOCUSS, to determine the latency and source of neuronal activity that is unique to the interference condition.
5. Each subject may be asked to have an MRI scan performed in the MRI facility located at Henry Ford Hospital. The MRI scans will be used to co-register the MEG data to specific locations in the cortex of each individual. This will allow for a precise localization of the anatomical landmarks and cortical activation areas associated with the control of attention and other executive controls over behavior. These locations will be used as a basis for comparison of each individual’s location of cortical functioning.
Data Analysis:
Single ECD [14,15]:
ECD will be used to determine the location, and amplitude of a brain electric activity for latencies where the MEG data exhibits peak power in channels over the occipital cortex during visual processing. The MSI ECD imaging software utilizes a gradient search algorithm with multiple seed points to determine the global optimum location and orientation of the dipole source that best accounts for the measured magnetic field pattern. In these calculations, the head will be modeled as a homogeneous sphere fit to the local skull surface geometry for each detector channel. The ECD results will be overlaid on the subject’s MRI scan.
MR-FOCUSS [31,32]:
The MR-FOCUSS technique produces a time sequence of whole brain images of both focal and extended source structures that are simultaneously active. For our proposed studies, the MEG and MRI data will be co-registered and a cortical model consisting of 2800 to 3000 cortical source location will be constructed such that each cortical model location represents an equivalent amount of cortical gray matter, as identified in the MRI data. The gain matrix for the cortical model will be constructed from forward model calculations for the MEG sensor array with separate X, Y, and Z oriented current dipoles at each cortical location. These calculations will employ a spherical head model with the radius and center matched to the local geometry of the skull derived from the MRI data. This technique images compact sources by using the FOCUSS [32] algorithm. However, it avoids sensitivity to noise by employing a multi-resolution mathematical model of the cortical source structure, as noise, which is uncorrelated with the cortical model magnetic fields, can be iteratively segregated from the data. This noise correlation minimization replaces the minimum norm technique in the FOCUSS iterative algorithm. The MR-FOCUSS results will be displayed on the subject’s MRI scan. At each latency, an initial estimate of the cortical model source amplitudes is refined until the best fit to the data is achieved. For these studies, an initial estimate of cortical structure is unavailable. Therefore, the X, Y, and Z source amplitudes of each of the cortical model locations will be initialized with a random number generator that will produce amplitudes characterized by a mean of zero and a standard deviation of 1 nanoamp-meter. At each iterative step, both the amplitude and orientation of the individual sources are altered. For statistical robustness, twenty of these solutions will be generated and averaged for each time slice of MEG data. Because of differences in initialization each solution is slightly different such that the average accentuates those features that are common to the set of solutions.
For each subject the latency (in ms), location (x,y,z coordinates) and average amplitude of response (nAm/time point) will be extracted from the MR-FOCUSS imaging results for each cognitive process step. This will be accomplished using a mathematical technique that identifies local source amplitude maxima within the volume of the cortical model. The identified cortical regions of interest will include: visual cortex (Brodmann’s areas 17-19), dorsal region of right inferior parietal cortex (Brodmann’s area 39/von Bonin & Bailey’s areas PG & OPT), ventral region of right inferior parietal cortex (Brodmann’s area 39 & 40/von Bonin & Bailey’s areas PG & PFG & PF), dorsal region of left inferior parietal cortex (Brodmann’s area 39/von Bonin & Bailey’s areas PG & OPT), dorsal region of right and left dorsolateral cortex (Brodmann’s areas 6,8 & 46/von Economo’s areas FB, FC & FD), ventral region of right dorsolateral frontal cortex (Brodmann’s areas 6,8 & 46/von Economo’s areas FB, FC & FD), right and left anterior cingulate cortex, (Brodmann’s area 24)/von Economo’s area LA), and right and left posterior cingulated cortex, (Brodmann’s area 23/ von Bonin & Bailey’s area LC). Activation will be studied for these regions as well as in the cerebellum and caudate nucleus of the basal ganglia. Quantified activation values will be used to perform statistical analysis.
Neuropsychological Measures:
Comparison of location and activation strength (nAmp-meters) of recorded MEG cortical activation and test scores for ADHD subjects will be compared to control subjects test scores and MEG activation to determine if MEG can be used in the determination of cortical areas involved in cognitive functioning differences between subjects with ADHD and controls. This comparison will determine if the amplitude of the cortical activation is correlated to the scores determined in the Wender-Utah Rating Scale and the Conner’s Adult ADHD Rating Scale .
The Wender-Utah Rating Scale and the Conner’s Adult ADHD Rating Scale will be used to diagnose the adults tested in the proposed study. This instrument will provide the following subtype diagnoses to be used in the testing of specific hypotheses: Attention Deficit – Hyperactive Disorder, Predominantly Inattentive Type, Attention Deficit – Hyperactive Disorder, Predominantly Hyperactive-Impulsive Type, and Attention Deficit – Hyperactive Disorder, Combined Type. These diagnostic/behavioral data will be used to form subgroups that will test the specific hypotheses regarding differences between the clinical subjects and normal controls subjects in the activation of brain regions. Due to the small number of subjects that will be supported by this pilot work, we will limit our testing to subjects that are predominately Inattentive or Hyperactive.
Subject group comparisons will be made of the location and activation strength (nAmp-meters) of the recorded MEG to determine whether or not there are differences in the cerebral regions involved in the control of attention and other executive functions as measured by different conditions in the Continuous Performance Test and the Stroop Test.
The latencies of MEG activation between the subject groups will also be compared for similarly-activated cerebral regions to determine whether or not there are differences in the efficiency of particular processes. Prolonged latencies will be interpreted as a reflection of a decrease in the efficiency of processing within the preceding activated region.
Additional MEG solutions will be viewed as a .gif movie that displays the millisecond-by-millisecond cerebral activation that occurs during each processing stage involved in the control of attention and other executive functions. The localization of cerebral activity occurring at different processing latencies will be used to compare the normal stages of processing (Normal Control Subjects) with the processing that is found for the subtypes of ADHD (ADHD Inattentive Type, ADHD Hyperactive/Impulsive Type). This analysis will allow us to remain open to additional hypotheses as we seek a better understanding of the neural pathways involved in both normal and abnormal attention and executive control.
The localization and amplitude of the evoked MEG field response during the performance of tasks that demand the engagement of attention and executive control (Continuous Performance Test (CPT) and Stroop Interference Test) will be identified by MR-FOUCSS. The extent to which specific cerebral regions are involved during these tasks will be quantified using the amplitude and latencies of the evoked MEG responses.
Five adults with the ADHD (subtype diagnoses will be considered and an attempt will be made to balance the testing of those with Inattentive and Hyperactive ADHD – 2:3 or 3:2) and five adult control subjects will be tested in this pilot study. Student t-tests will be used to make pair-wise comparisons between the normal control and ADHD subtypes for each of the hypothesized results. Only very large effect sizes of 1.6 will be detected with 80% power. Each pair-wise comparison will be performed at the 0.05 level since each comparison is pre-specified. If the underlying t-test assumptions are significant violated than the nonparametric Wilcoxon rank sum test will be used for analysis of these initial data.
Significance of the proposed study:
MEG systems providing whole head coverage are available for clinical application as well as basic neuroscience studies. There are currently 100 such systems known to be in use worldwide, but only 12 whole head systems in the United States . In this country, the major clinical application of MEG has been pre-surgical and epilepsy mapping. Due to the high cost of an MEG system and shielded room, additional clinical applications are needed to make MEG cost-effective. In this proposal whole head MEG will be used to obtain a better understanding of the stages involved in the control of attention and other executive functions in normal control subjects and patients diagnosed with attention-deficit hyperactivity disorder (ADHD) [33, 34].
ADHD is the most common psychiatric disorder of childhood that continues into adulthood for approximately 30% of the cases. Attention to this public health problem has increased in recent years. In 1998, the National Institutes of Health held a consensus development conference on the diagnosis and treatment of ADHD [http://consensus.nih.gov/cons/110/110_statement.htm]. The conclusions of this panel were briefly: 1) ADHD poses a costly public health threat, 2) no reliable method of diagnosis exists, 3) it is a disorder of multiple dimensions, 4) long-term randomized, medication clinical trials are needed, 5) medication use varies widely, 6) diagnosis, treatment, and follow-up need improvement, 7) causes need to be understood to develop prevention strategies. In 1999, the Developmental Disabilities Branch of the Center of Disease Control and Prevention sponsored a conference on ADHD [http://www.cdc.gov/ncbddd/adhd/], and in 2000, the American Academy of Pediatrics released a set of guidelines for evaluation [http://www.aap.org/policy/ac0002.html]. These guidelines seek to distinguish children whose condition warrants treatment from those exhibiting behavior that is developmentally normal. This proposed study should help to clarify the causes of ADHD that would lead to the development of more effective treatment strategies.
This proposed study will ascertain whether MEG can be used to detect cortical activation differences in subjects with ADHD compared to control subjects. Current Density mapping by MR-FOCUSS will provide a MEG imaging technique capable of detecting the specific cortical areas and the neuronal pathways involved in cognitive processes. Millisecond temporal resolution of these MR-FOCUSS results will provide additional information on how the ADHD brain is processing higher order cognitive functioning. The results from MEG imaging will provide a more integrated image of neuronal processing during cognition. There is currently no other imaging technique available providing combined temporal resolution and high spatial resolution in a safe, non-invasive imaging modality. There is every reason to believe that MEG may provide a valuable diagnostic tool for ADHD and other neuropsychological disorders.
References
1. Nagel-Leiby, S., Buchtel, H.A., Welch, K.M.A. (1990). Cerebral control of directed visual attention and orienting saccades. Brain, 113:237-276, 1990.
2. Berquin, PC, Geidd, JN, Jacobsen, LK, Hamburger, MA, Krain , AL , Rapoport, JL, Castellanos, FX. (1998). Cerebellum in attention-deficit hyperactivity disorder: A morphometric MRI study. Neurology, 50, 1087-1093.
3. Brandeis, D, Banaschewski, T, Baving, L. (2002). Multicenter P300 brain mapping of impaired attention to cues in hyperactivity children. Journal of the American Academy of Child and Adolescent Psychiatry, 42(8) , 990-998.
4. Coull, JT, Nobre, AC. (1998), When and where to pay attention: The neural systems for directing attention to spatial locations and to time intervals as revealed by both PET and fMRI. The Journal of Neuroscience, 18(8), 7426-7435.
5. Driver, J, Frackowiak, RSJ. (2001). Neurobiological measures of human selective attention. Neuropsychologia, 39, 1257-1262.
6. Fernandez-Duque, D, Baird, JA, Posner, MI. (2000). Executive attention and metacognitive regulation. Consciousness and Cognition, 9, 288-307.
7. Gruber, SA, Rogowska, J, Holcomb, P, Soraci, S, Yurgelun-Todd. (2002). Stroop performance in normal control subjects: An fMRI study. Neuroimage, 16, 349-360.
8. Geidd, JN, Blumenthal, J, Molloy, E, Castellanos, FX. (). Brain imaging of attention deficit/hyperactivity disorder. Annals New York Academy of Science , 33-49.
9. Kanwisher, N, Wojciulik, E. (2000). Visual attention: Insights from brain imaging, Nature Reviews Neuroscience, 1, 91-100.
10. MacLeod, CM, MacDonald, PA. (2000). Interdimensional interference in the Stroop effect: Uncovering the cognitive and neural anatomy of attention. Trends in Cognitive Sciences, 4(10), 383-391.
11. Overmeyer, S, Taylor, E. (2001). Neuroimaging in hyperactivity children and adults: an overview. Pediatric Rehabilitation, 4(2), 57-70.
12. Ricco , CA , Reynolds, CR, Lowe, P, Moore, JJ. (2002). The continuous performance test: A window on the neural substrates for attention?, Archives of Clinical Neuropsychology, 17, 235-272.
13. Nakamura A, Kakigi R, Hoshiyama M, Koyama S, Kitamura Y, Shimojo M (1997). Visual Evoked Cortical Magnetic Fields to Pattern reversal stimulation. Cognitive Brain Research 6, 9-22.
14. H ä m ä l ä inen M, Hari R, Ilmoniemi J, Knuutila J, Lounasmaa OV (1993). Magnetoencephalography-theory, instrumentation, and applications to noninvasive studies of the working human brain. Review of Modern Physics, 65(2) , 413-497.
15. Sato S, (ed) (1990). Magnetoencephalography, Raven, New York .
16. Moran, J.E., Tepley, N., (2000). Two dimensional inverse imaging (2DII) of current sources in magnetoencephalography. Brain Topography, 12 ,201-217.
17. Fogelson, DL, Nuechterlein, KH, Asarnow, RF, Subotnik, KL, Talovic, SA., (1991). Interrater reliability of the structured clinical interview for DSM-II-R, Axis II: schizophrenia spectrum,m and affective spectrum disorders. Psychiatry Research, 39(1), 55-63.
18. Folstein, MF, Folstein, SE, McHugh, PR., (1975). Mini-mental state: A practical method for grading the cognitive state of patients for the clinician. J. Psychiatric Research, 12(3), 189-193.
19. Ward, M.F., Wender, P.H. & Reinherr, F.W., (1993). The Wender Utah Rating Scale: An aid in the retrospective diagnosis of childhood attention deficit hyperactivity disorder, American Journal of Psychiatry, 150(6), 885-890.
20. Stein, M.A., Sandoval, R., Roizen, M.D., Reincke, M.A., Blondis, T.A. & Klein, Z., (1995). Psychometric characteristics of Wender Utah Rating Scale (WURS): Reliability and factor structure for men and women, Psychopharmacological Bulletin , 31(2) , 425-433.
21. Conners, CK., (1994). The Conners Rating Scales: Use in clinical assessment, treatment planning and research, In M. Maruish (ed.), Use of psychological testing for treatment planning and outcome assessment, Hillsdale , NJ : L. Erlbaum.
22. Ramanaiah , NV , Franzen, M, Schill, T. (1983). A psychometric study of the State-Trait Anxiety Inventory, J Personality Assessment, 47(5), 531-535.
23. Epstein, J.N., Goldberg, C.K., Conners, C.K. & March, J.S., 1997. The effects of anxiety on continuous performance test functioning in an ADHD clinic sample, Journal of Attention Disorders, 2(1), 45-52.
24. Cloninger, C.R. 1987. A systematic method for clinical description and classification of personality variants, Archives of General Psychiatry, 44, 573-588.
25. Cloninger, C.R., Adolfsson, R., Svrakic, D.M., 1996. Mapping genes for human personality, Nature – Genetics, 12, 3-4.
26. Cloninger, C.R., Frzybeck, T.B. & Svrakic, D.M., 1991. The tridimensional personality questionnaire: U.S. normative data, Psychological Reports, 9, 1047-1057.
27. Cloninger, C.R., Svrakic, D.M., Przybeck, T.P. 1993. A psychobiological model of temperament and character, Archives of General Psychiatry, 50, 975-990.
28. Heath, A.C., Cloninger, C.R. & Martin, N.G. 1994. Testing a model for the genetic structure of personality: A comparison of the personality systems of Cloninger and Eysenck, Journal of Personality and Social Psychology, 66(4), 72-775.
29. Rosvold, H, Mirsky, A, Sarason, I, Bransome, ED, Beck LH. (1956). A continuous performance test of brain damage. Journal of Consulting Psychology, 30, 343-350.
30. Stroop, LR, (1935). Studies of interference in serial verbal reactions. Journal of Experimental Psychology , 28, 643-662.
31. Moran, J. (2003) MR-FOCUSS. 2DII user manual. Available for download from: http://rambutan.phy.oakland.phy.edu/~meg.
32. Gorodnitsky, IF, George, JS, Rao, BD. (1995). Neuromagnetic source imaging with FOCUSS: a recursive weighted minimum norm algorithm. Electroencephalography and Clinical Neurophysiology, 95, 231-251.
33. Filipek, P. (1999). Neuroimaging in the developmental disorders: The state of the science. Journal of Child Psychology and Psychiatry. 40(1). 113-128.
34. Rapoport, JL. Castellanos, X, Gogate, N, Janson, K, Kohler, S, Nelson, P. (2001). Imaging normal and abnormal brain development: New perspectives for child psychiatry. Australian and New Zealand Journal of Psychiatry , 35, 272-281.