International Journal of Psychophysiology 53 (2004) 207 – 216 www.elsevier.com/locate/ijpsycho
Specific EEG frequencies signal general common cognitive processes as well as specific task processes in man Thalı´a Harmony a,*, Thalı´a Ferna´ndez a, Jorge Gersenowies b, Lı´dice Gala´n c, Antonio Ferna´ndez-Bouzas a, Eduardo Aubert c, Lourdes Dı´az-Comas c a
Instituto de Neurobiologı´a, Campus UNAM-UAQ Juriquilla, Apartado Postal 1-11141 Quere´taro, Qro, 76230, Mexico b FES Iztacala, UNAM, Los Reyes, Tlanepantla, Estado de Me´xico, Mexico c Centro de Neurociencias de Cuba, Playa, La Havana, Cuba Received 13 May 2003; received in revised form 30 March 2004; accepted 7 April 2004 Available online 15 June 2004
Abstract The EEG of 10 normal male young adults was recorded during the performance of three different tasks: mental calculation, verbal working memory (VWM) and spatial working memory (SWM). The stimuli used in the three tasks were the same, only the instructions to the subjects were different. Narrow band analysis of the EEG and distributed sources for each EEG frequency were calculated using variable resolution electromagnetic tomography (VARETA). At some frequencies (1.56, 4.68, 7.80 to 10.92 Hz) at least two tasks produced similar EEG patterns that were interpreted as the reflex of common cognitive processes, such as attention, inhibition of irrelevant stimuli, etc. Specific changes were also observed at 2.34, 3.12, 3.90, 5.46 and 6.24 Hz. The first three of these frequencies showed similar changes during VWM and calculus at the left frontal cortex, suggesting the activation of working memory (WM) processes. The interaction effect at these frequencies was mainly observed at the anterior cingulate cortex and frontal cortex. At 5.46 and 6.24 Hz, changes were only observed during mental calculation. D 2004 Elsevier B.V. All rights reserved. Keywords: EEG; Mental tasks; Frequency analysis; qEEG; Source analysis; Cognitive processes; VARETA
1. Introduction In previous papers, it was proposed that specific frequencies may signal specific cognitive processes during mental tasks. This proposal was based on results obtained during mental calculation as opposed to a control experimental condition (Harmony et al., 1999), during the execution of a verbal working * Corresponding author. Tel.: +52-55-5623-4051; fax: +52-555623-4005. E-mail address:
[email protected] (T. Harmony). 0167-8760/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.ijpsycho.2004.04.006
memory (VWM) task in young adults and children (Ferna´ndez et al., 2002) and during semantic categorization of figures and words in children (Harmony et al., 2001). If this hypothesis is correct, then during the execution of different tasks triggered by the same stimulus, those processes common to the three tasks should activate the same frequencies in the same regions, and those processes specific to a particular task should activate different frequencies or the same frequencies in different regions. In this work, we addressed this issue by having subjects perform three different tasks triggered by the
208
T. Harmony et al. / International Journal of Psychophysiology 53 (2004) 207–216
same stimuli. The tasks were a verbal working memory (VWM) task, a spatial working memory (SWM) task and an arithmetic addition (calculus) task. According to the model of Baddeley and Hitch (Baddeley, 1999), three different components may be distinguished in working memory (WM): the central executive (CES), the articulatory loop and the visuospatial scratch pad. The articulatory loop comprises two subsystems: an active subvocal rehearsal process and a passive, phonologically based storage component. The rehearsal process is the form of subvocal articulation, and it is closely linked with speech production. The contents of the passive store are subject to decay, but they can be maintained by subvocal rehearsal. An important aspect is that the contents of WM do not stay in memory for long, unless constant attention is given to them. Thus, sustained attention to internal processing is a very important element during mental calculation and may be considered either a component independent from WM or as a part of it (Baddeley, 1999). The tasks used in this experiment were designed so that they shared common processes such as the activation of CES and differed in one or more processes. SWM will activate the visuospatial scratch pad and VWM the articulatory or phonological loop. Mental calculation will include VWM processes as well as other processes, such as assignation of magnitudes to numerical quantities, mental calculation per se, storage of intermediate results, retrieval of arithmetic facts from long-term memory, etc. We analyzed the EEG segments during two different intervals on each trial: ‘‘pre-segments’’ were selected 1280 ms prior to the presentation of the memory set and ‘‘post-segments’’ 1280 ms prior to the presentation of the probe (Fig. 1). It has been shown that the most significant differences between tasks are observed when the differences between the
logarithm of the power of the EEG during task performance and the logarithm of the power of the EEG segment prior to the stimulus are compared for each task (Ferna´ndez et al., 1995). Thus, we are evaluating the rate of change between the EEG during task performance and the EEG before the stimulus for each subject. On the other hand, if we analyze the differences in the psychological conditions to which the subject is exposed for each interval, we can suppose that after the warning stimulus and prior to the presentation of the memory set, the subject is in a state of alertness and expectancy waiting for the stimulation. After the presentation of the memory set, he is also in a state of alertness and expectancy waiting for the probe stimulus in the three tasks. However, according to the instructions, he is also performing the specific task. In the VWM task, the subject is maintaining the memory set in the temporal phonological storage of the verbal working memory by rehearsal in the articulatory loop. In the SWM task, he is maintaining the spatial characteristics of the memory set in the temporal storage of the working memory by rehearsal of the spatial code (Luria, 1978; Jonides and Smith, 1997). During calculation, the subject is maintaining in the temporal phonological storage the memory set as well as the partial results and the final result, and he accesses and retrieves from long-term memory storage the arithmetic facts necessary to perform the addition (McCarthy and Warrington, 1990; McCloskey, 1992; Hitch, 1978). During the performance of these tasks, there are also other processes common to the three tasks: recognition of numbers in their Arabic form, comprehension of verbal representation of numbers and all processes assigned to the central executive system (CES). According to Jonides and Smith (1997), CES may be integrated by a constellation of processes or by a set of executive processes
Fig. 1. Time chart for the tasks. EEG segments (pre-segments) were selected 1280 ms prior to the presentation of the memory set and 1280 ms prior to the presentation of the probe (post-segments). WS, warning stimulus.
T. Harmony et al. / International Journal of Psychophysiology 53 (2004) 207–216
that are activated depending on the task. These authors considered that one process is attentional, focusing attention on one part of the problem, another process is inhibition, since in many tasks, it is necessary to inhibit previous information that was the focus of attention so that current processing can focus on new information and other processes are scheduling operations and setting priorities for different tasks.
2. Materials and methods The protocol was approved by the Ethics Committee of the Institute of Neurobiology. Subjects were told about the procedure, and they gave their informed consent. The EEG was recorded of 10 male right-handed volunteers (20 – 26 years old) without neurological antecedents and with normal EEGs. For the three tasks, the same memory set and time chart were used (Fig. 1). Each trial in this experiment began with a visual warning stimulus (*) lasting 300 ms. After an interval of 2 s, a set of five digits (memory set) was presented on a video monitor for 1500 ms, and 2 s later, a probe stimulus was displayed for 300 ms. The interval from end of probe to next warning stimulus was 3 s. Two hundred trials were presented. The tasks were: (1) Verbal working memory task (VWM): We used a modification of Sternberg’s (1966) paradigm. In this task, the probe was a single digit. The subject had to respond with one button if the probe was in the memory set and with another button if it was not. In 50% of the trials, the number belonged to the memory set. (2) Spatial working memory (SWM): In this task, the probe was a digit within a sequence of dashes and spaces (i.e., - - 5 - -). The subject has to remember the position of the different digits in the memory set and to respond with one button if the position of the probe was the same as in the memory set, or with another button if it was not. (3) Calculus: In this task, the probe was a number. The subject has to add the numbers of the memory set and to respond with one button if the number shown in the probe stimulus agreed with the result of the addition and with another button if it did not. In
209
50% of the trials, the probe gave a correct addition. The possible solutions were numbers between 10 (0 + 1 + 2 + 3 + 4) and 35 (5 + 6 + 7 + 8 + 9), and for this reason, the probe was within this range. Furthermore, to be sure that subjects performed the task, if in the memory set there was an even number of odds, the probe was an even number, and if, on the contrary, there was an odd number of odds, the probe was an odd number. In all tasks, subjects used both hands (counterbalanced) in go-go paradigms. The order of presentation of the tasks was also counterbalanced across subjects. The number of correct responses and reaction times were measured. The EEG was recorded with reference to linked ears from Fp1, Fp2, F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T3, T4, T5, T6, Fz, Cz, Pz and Oz. The EOG was recorded from a supraorbital electrode and from an electrode placed in the external canthus of the right eye. Recording conditions have been described elsewhere (Harmony et al., 1996). As stated previously, in all tasks, EEG segments of 1280 ms were selected immediately before the memory set (pre-segments) and immediately before the presentation of the probe stimulus (post-segment) (Fig. 1). These segments were visually edited for the selection of artifact-free segments. Only correct responses were analyzed. 2.1. Data analysis Each of these segments of EEG recording is a vector-valued time series that comprises Nd = 19 components, one for each derivation. This vector, for a given time instant t, shall be denoted as Vi,jraw(t), where i varies from 1 to Ns (number of subjects), j from 1 to Nseg (number of EEG segments), and t varies from 1 to Nt = 128(number of time instants). Each Vi,jraw(t) was transformed to the frequency domain by means of the Fast Fourier Transform (FFT), producing a set of complex valued vectors Vi,jraw(x), where x indexes frequency and varies from 1 to Nx = 24, the number of frequencies used for subsequent analyses. These indices correspond to the frequencies from 0.78 to 19.11 Hz, every 0.78 Hz. Spectra and cross-spectra are essentially measures of the variance and covariance, respectively, of complex valued Fourier coefficients. They contain all the
210
T. Harmony et al. / International Journal of Psychophysiology 53 (2004) 207–216
relevant information in the case of stationary Gaussian linear signals.
specified as follows in the frequency domain (Casanova et al., 1996):
2.2. EEG source spectral analysis
Vi ðxÞ ¼ K Ji ðxÞ þ Ei ðxÞ
Variable resolution electrical tomography (VARETA) is a recently developed technique (Valde´s et al., 1996; Valde´s-Sosa et al., 1996; Bosch-Bayard et al., 2001) for estimating the source generators of EEG data. It is a discrete spline distributed solution (Riera et al., 1997). Spline estimates are the spatially smoothest solutions compatible with the observed data. The first type of discrete spline solution describe was LORETA (Pascual-Marquı´ et al., 1994; Pascual-Marquı´, 1994, 1995), which is simple to implement and yet localizes point sources with great accuracy. Since LORETA imposes maximal spatial smoothness, it estimates distributed solutions that are already smooth with little error. However, point sources are blurred. VARETA, on the other hand, imposes different amounts of spatial smoothness for different types of generators selected by a datadriven search procedure. In addition, in VARETA, current sources are restricted to gray matter by use of a probabilistic mask that prohibits solutions where the mask is zero, i.e., in CSF or white matter. The Probabilistic brain atlases (PBA) used was developed at the Montreal Neurological Institute (Evans et al., 1993, 1994; Collins et al., 1994; Mazziotta et al., 1995). The mean head used in this work was obtained (Evans et al., 1994) by averaging a set of 305 normal MRI scans transformed to Talairach space. The MRIs had been subjected to nonlinear warps to match a set of 50 common landmarks. Use of the PBA is intended to provide approximate solutions. The three concentric sphere model was fitted to the MNI mean head by means of a least-square algorithm. The brain electric source analysis (BESA, Scherg and Von Cramon, 1985) coordinates for each electrode were then used to project each electrode onto the average skin. For the head volume conductor model, it was assumed that the conductivities of the three spheres were 1.0 for the skin, 0.0125 for the skull and 1 for the brain. The ratio of the radii of the spheres was standardized to 1 for the skin, 0.94 for the skull and 0.86 for the brain. Based on this head model, the forward problem of EEG may be
where the magnitudes V(x) and Ei(x) denote the data and error vectors measured on the sensor array described in this paper. Ji (x) represents the matrix of the x, y and z components of the primary current field discretized on a grid inside the brain. In this work, the grid was a set of size Ngrid = 3623. K is the lead field that relates current densities to observed measurements obtained by discretization (Riera et al., 1997). A complete treatment of the problem would require a discussion of the reference. Following Pascual-Marquı´ et al. (1994), we transform the data and the matrix K to the average reference. The usual objective of distributed EEG inverse solutions is to estimate the primary current field Ji from the data matrix Vi. This is the inverse EEG problem for the source time series. On the other hand, source EEG spectral analysis has a different objective: the estimation of S(x) the source cross-spectrum. This procedure, called frequency-domain VARETA, has been applied to the localization of brain tumors, infarcts and other type of cerebral lesions with great accuracy, including subcortical lesions (Ferna´ndez-Bouzas et al., 1999, 2000, 2001, 2002; Prichep et al., 2001). 2.3. Statistical analysis Special purpose 3D graphical tools were developed in order to allow interactive evaluation of the large amount of transformed data that are available. Statistical analysis (Statistical Parametric Mapping) was performed, using each voxel (point of the grid) as a dependent variable. Correction for multiple comparisons was done (Worsley et al., 1995). In a generalization of topographic maps, three-dimensional colorcoded images were generated, in which the color code reflects the value of a given voxel in the statistical test performed. Extreme values will show up as ‘‘hot spots’’ as has become standard in Statistical Parametric Mapping (SPM). Using the values of all subjects resulting from the source analysis at each point of the grid (voxel) at each frequency, a repeated-measures multivariate ANOVA was carried out for the following factors: task (VWM, SWM, calculus) and EEG segment (prestimuli vs. poststimuli). Post hoc Stu-
T. Harmony et al. / International Journal of Psychophysiology 53 (2004) 207–216 Table 1 Behavioral results Task
VWM SWM Calculus
211
Table 3 Interaction effects
Number of correct responses (200 trials)
RT
Frequency (Hz)
Mean
S.D.
Mean
S.D.
179.56 158.55 153.56
5.20 5.27 7.12
955.33 1234.67 862.22
67.67 54.40 77.84
dent’s t-tests between pre- and post-segments for each task were also calculated.
F*
2.34
5.78
3.12
6.96
3.90
9.11
5.46 6.24
12.88 6.30
Topography (F) bilateral superior frontal, anterior cingulate dorsolateral frontal, anterior cingulate right frontal, anterior cingulate all cortex left premotor region, anterior cingulate
F452 > 5.11, p < 0.01.
3. Results 3.1. Behavioral results Mean and standard deviations of number of correct responses and reaction times are shown in Table 1.
ANOVA between tasks showed that the number of correct responses was greater for SWM and calculus than for VWM ( F = 5.41, p < 0.0115). Reaction time was greater for SWM than for VWM and calculus ( F = 8.29, p < 0.0018).
Table 2 EEG segment effects Rm ANOVA
t-tests
Frequency (Hz)
F*
Topography (F)
Calculus (t)
VWM (t)
SWM (t)
1.56 2.34
38.19 31.49
frontal, temporal all cortex
z frontal z all cortex
z frontal z right occipital
3.12
42.69
all cortex
z all cortex
3.90
49.57
all cortex
z all cortex
4.68
39.71
5.46
22.98
z frontal, parietal, cingulate z all cortex
6.24
12.53
z parietal, cingulate
NS
NS
7.02 7.80 8.58
13.26 20.71 18.67
all cortex except occipital frontal, anterior temporal frontal, anterior temporal occipital all cortex all cortex
z right prefrontal z prefrontal, left anterior cingulate z left inferior frontal, bilateral occipital z left prefrontal, anterior cingulate z frontal, parietal, cingulate NS
NS # right frontal NS
NS NS NS
9.36
25.60
all cortex
# left occipital
NS # prefrontal # bilateral dorsolateral frontal, cingulate # right inf. temporal
10.14
26.12
all cortex
# occipital
NS
10.92 13.28
21.50 7.60
all cortex left temporal
# left temp-occipital NS
# left occipital left occipital
F451 > 7.23, p < 0.01. z current increases at post-segments ( p < 0.01). # current decreases at post segments ( p < 0.01).
z bilateral occipital, right temporal NS NS NS
# bilat. superior frontal gyri, anterior cingulate # right parietal, right superior temporal NS NS
212 T. Harmony et al. / International Journal of Psychophysiology 53 (2004) 207–216
Fig. 2. T values for the differences between post-pre EEG segments. Yellow indicates significance, p < 0.01. (A) SWM. (B) VWM. (C) Calculus. (D) F values for the interaction task segment, yellow indicates significance, p < 0.01. The level of the slide is shown at the right. The lines on the images at the right indicate the brain section.
T. Harmony et al. / International Journal of Psychophysiology 53 (2004) 207–216
3.2. Electrophysiological results The task effect, independent of the segment, was significant ( p < 0.01) at 1.56, 3.90, 5.46 and 7.02 Hz in Broca’s area, left prefrontal region, thalamus and right inferior frontal gyrus, respectively. This task effect was also observed at 9.36 –10.92 Hz in occipital areas and at 13.28 and 14.82 Hz in the right angular gyrus. The segment effects are shown in Table 2. F-values and the areas where they were significant ( p>0.01) are presented in the first two columns. The last three columns showed the regions where the t-values between post- and pre- segments were significant ( p < 0.01) for each task. F-values were significant in many frequencies and in wide regions of the cortex. In each task, the small arrows indicate whether the current increased or decreased in the post-segment, according to the t-value. Table 3 shows the F-values obtained for the interaction with rmANOVA. The interaction effect, or the differences between tasks in the change from pre- to post-segments, was significant at 2.34, 3.12, 3.90, 5.46 and 6.24 Hz. For these five frequencies, post hoc analysis showed that in the mental calculation task the t-values were all significant. The interaction was significant at 2.34 Hz in bilateral superior frontal and anterior cingulate regions, at 3.12 Hz in both dorsolateral frontal and anterior cingulate regions, at 3.90 Hz in the right frontal lobe and anterior cingulate region, at 5.46 Hz in all of the cortex mainly in the right hemisphere and at 6.24 Hz in the left premotor region and anterior cingulate area. Fig. 2 shows the distribution of the F-values at 2.34 Hz (D), as well as the distribution of t-values between segments for each task. At this frequency, all the cortex showed significant increases of the current during calculus (C). Frontal areas, particularly in the left hemisphere, showed significant changes during VWM (B). During SWM (A), the current increases at this frequency were observed at the right occipital region. Interaction effects (D) may be observed at anterior cingulate cortex and frontal cortex. At 3.12 Hz, all of the cortex showed current increases during mental calculation, current increases in the left inferior frontal and right occipital regions during VWM and in bilateral occipital and right temporal regions during SWM. At 3.90 Hz, signifi-
213
cant changes were observed in all of the cortex for calculus, in left frontal and anterior cingulate during VWM, and no significant changes were observed during SWM. At 5.46 and 6.24 Hz where a strong interaction effect was shown, significant changes from pre- to post-segment were observed only during mental calculation in all of the cortex and at cingulate and bilateral parietal superior areas, respectively.
4. Discussion 4.1. Behavioral results VWM was the task with the most correct responses, suggesting that this task was easier than the other two. Reaction time was smaller in calculus and VWM, supporting the conclusion that SWM was the most difficult task. 4.2. EEG results VWM is a task included in mental calculation; thus, we expected common activation of the structures related to VWM in these two tasks (Hitch, 1978). As observed in Table 2, from 1.56 to 4.68 Hz, frontal cortex shows current increases at the post-segment in these two tasks. At 1.56 Hz, the SWM task also activated the frontal cortex. Common processes to the three tasks are those related with CES. Baddeley (1999) proposed that focusing attention is one of the main functions of the CES. In this case, we may suppose that sustained attention to internal processing is one of the common processes. Harmony et al. (1996) suggested that the appearance of delta activity in frontal regions may indicate attention to internal processing during performance of mental tasks. A delta increase has been reported in some experiments during mental arithmetic or other types of mental tasks (Dolce and Waldeier, 1974; Kakizaki, 1984; Tucker et al., 1985; Etevenon, 1986; Ferna´ndez et al., 1995). Valentino et al. (1992) observed an increase in delta power in anterior regions between rest and a continuous performance task condition. They do not preclude the possibility that eye movements were responsible, though in a pilot study of 10 subjects, they did not find any consistent relationship between EOG and frontal power. In our case, we carefully
214
T. Harmony et al. / International Journal of Psychophysiology 53 (2004) 207–216
rejected the EEG segments with eye movements; thus, we consider it very unlikely that the increase in delta is due to ocular artifacts. Segment effects for the three tasks were also observed in the whole cortex for frequencies 7.8, 8.58, 9.36, 10.14 and 10.92 Hz, but no interaction effects were observed for these frequencies within the alpha band, indicating that the changes at these frequencies are more directly related to generalized processes involved in the three tasks than to any specific process. Alpha suppression during mental tasks has been a common observation since EEG was first recorded (Adrian and Matthews, 1934). According to Klimesch (1999), desynchronization in the lower alpha band is observed in response to several unspecific factors and is topograpically widespread, probably reflecting general task demands such as attentional processes. Only frequencies at 2.34, 3.12, 3.90, 5.46 and 6.24 Hz showed significant interactions. In general, for all frequencies, the interaction effect between segments and tasks was observed in frontal and anterior cingulate regions. In the first three of these frequencies during calculus, there was a generalized increase of current, whereas in VWM, this increase of current was observed mainly in the left frontal cortex, and no changes were observed in SWM. Common processes activating frontal structures in VWM and calculus are those related to the articulatory processes of the phonological loop. In a previous paper on mental calculation, we reported that 3.90 Hz signaled such processes (Harmony et al., 1999), and as we have mentioned, delta activity has been referred during mental calculation (Dolce and Waldeier, 1974; Kakizaki, 1984; Etevenon, 1986). The other structure where interaction was shown was the anterior cingulate cortex. This structure was assigned the role of the attentional executive processes by Posner (1988). According to Mesulam (1990), the anterior cingulate provides the emotional value during attentional processes. Recently, Mulert et al. (2003) have reported a relationship between activation of anterior cingulate cortex and behavioral results. Increased activity of the anterior cingulate cortex is associated with an improvement in one aspect of performance (speed) and in the deterioration of another (accuracy). Our behavioral results showed that calculus, together with SWM, produced more incor-
rect responses and that calculus showed the smaller reaction time, which supported the results of Mulert et al. These authors propose that the anterior cingulate cortex is involved in performance monitoring, since increased speed would be expected to increase the probability of errors, requiring intensified performance monitoring. Also very interesting was the presence of occipital, but not of anterior changes during SWM at 2.34 and 3.12 Hz. Activation of the visuo-spatial loop is characterized by concomitant activation of the posterior cortex (Jonides and Smith, 1997), which may explain our results. However, we were not able to detect any specific activation of the right frontal cortex during SWM. This needs further experimental clarification. Studies reporting EEG activation during spatial tasks unfortunately have used very few electrodes, and the changes reported have been in the beta band (Gutierrez and CorsiCabrera, 1988). The interaction was also observed at 5.46 and 6.24 Hz in all of the cortex and in the left frontal lobe, respectively. According to Gutierrez and CorsiCabrera (1988) and Legewie et al. (1969), the most important differences between tasks are observed in the theta band. These frequencies increased at the post-segments only during mental calculation. Theta increases have been associated with attentive states (Gevins et al., 1997; Sasaki et al., 1996), episodic memory and the encoding of new information (Klimesch, 1999) and memory load (Gevins et al., 1997). Our findings support these results since of the three tasks, calculation involved long-term memory, access of new information and the greatest memory load. We conclude that specific EEG frequencies signal specific task processes as well as general common cognitive process in man.
Acknowledgements This project was partially supported by grants of CONACyT and PAPIIT, UNAM. The authors are grateful to Dr. Dorothy Pless for her revision of the English version and to engineer He´ctor Belmont, Ms. Rosa Marı´a Herna´ndez, Mr. Rafael Silva Cruz and Ms. Pilar Galarza for their technical assistance.
T. Harmony et al. / International Journal of Psychophysiology 53 (2004) 207–216
References Adrian, E.D., Matthews, B.H.C., 1934. The Berger rhythm: potential changes from the occipital lobes of man. Brain 57, 354 – 385. Baddeley, A., 1999. Human Memory. Allyn and Bacon, Boston, MA. 325 pp. Bosch-Bayard, J., Valdes-Sosa, P., Virues-Alba, T., Aubert-Vazquez, E., John, E.R., Harmony, T., Riera-Diaz, J., Trujillo-Barreto, N., 2001. 3D statistical parametric mapping of EEG source spectra by means of variable resolution electromagnetic tomography (VARETA). Clin. Electroencephalogr. 32, 47 – 61. Casanova, R., Valde´s, P., Garcı´a, F.M., Aubert, E., Riera, J., Korin, W., Lins, O., 1996. Frequency domain distributed inverse solutions. In: Wood, C. (Ed.), Proceedings of the Tenth Interna´ 96. Pergamon tional Conference on Biomagnetism, BIOMAG Press, Santa Fe, New Mexico, pp. 119 – 124. Collins, D.L., Neelin, P., Peter, T.M., Evans, A.C., 1994. Automatic 3D registration of MR volumetric data in standardized Tailarach space. J. Comput. Assist. Tomogr. 18, 192 – 205. Dolce, G., Waldeier, H., 1974. Spectral and multivariate analysis of EEG changes during mental activity in man. Electroencephalogr. Clin. Neurophysiol. 36, 577 – 584. Etevenon, P., 1986. Applications and perspectives of EEG cartography. In: Duffy, F.H. (Ed.), Topography Mapping of the Brain. Butterworth, Stoncham, pp. 113 – 141. Evans, A.C., Collins, D.L., Mills, S.R., Brown, E.D., Kelly, R.L., Peters, T.M., 1993. 3D statistical neuroanatomical models from 305 MRI volumes. Proc. IEEE—Nuclear Science Symposium and Medical Imaging Conference, 1813 – 1817. Evans, A.C., Collins, D.L., Neelin, P., MacDonald, D., Kamber, M., Marret, T.S., 1994. Three dimensional correlative imaging. Applications in human brain mapping. In: Thatcher, R., Hallet, M., Zeffiro, T., John, E.R., Huerta, M. (Eds.), Functional Neuroimaging. Academic Press, New York, pp. 145 – 162. Ferna´ndez, T., Harmony, T., Rodrı´guez, M., Bernal, J., Silva, J., Reyes, A., Marosi, E., 1995. EEG activation patterns during the performance of tasks involving different components of mental calculation. Electroencephalogr. Clin. Neurophysiol. 94, 175 – 182. Ferna´ndez, T., Harmony, T., Gersenowies, J., Silva-Pereyra, J., Ferna´ndez-Bouzas, A., Gala´n, L., Dı´az-Comas, L., 2002. Sources of EEG activity during a verbal working memory task in adults and children. In: Reisin, R.C., Nuwer, M.R., Hallet, M., Medina, C. (Eds.), Advances in Clinical Neurophysiology (Supplement to Clinical Neurophysiology, vol. 54. Elsevier, Amsterdam, pp. 269 – 283. Ferna´ndez-Bouzas, A., Harmony, T., Bosch, J., Aubert, E., Ferna´ndez, T., Valde´s, P., Silva, J., Marosi, E., Martı´nez-Lo´pez, M., Casia´n, G., 1999. Sources of abnormal EEG activity in the presence of brain lesions. Clin. Electroencephalogr. 30, 46 – 52. Ferna´ndez-Bouzas, A., Harmony, T., Ferna´ndez, T., Silva-Pereyra, J., Valde´s, P., Bosch, J., Aubert, E., Casia´n, G., Otero-Ojeda, G., Ricardo-Garcell, J., Herna´ndez, A., Santiago, E., 2000. Sources of abnormal EEG activity in brain infarctions. Clin. Electroencephalogr. 31, 165 – 169. Ferna´ndez-Bouzas, A., Harmony, T., Ferna´ndez, T., Ricardo-Gar-
215
cell, J., Casia´n, G., Sa´nchez-Conde, R., 2001. Cerebral blood flow and sources of abnormal EEG activity (VARETA) in neurocysticercosis. Clin. Neurophysiol. 112, 2281 – 2287. Ferna´ndez-Bouzas, A., Harmony, T., Ferna´ndez, T., Aubert, E., Ricardo-Garcell, J., Valdes, P., Bosch, J., Casia´n, G., SanchezConde, R., 2002. Sources of abnormal EEG activity in spontaneous intracerebral hemorrhage. Clin. Electroencephalogr. 33, 70 – 76. Gevins, A., Smith, M., McEvoy, L., Yu, D., 1997. High-resolution EEG mapping of cortical activation related to working memory: effects of task difficulty, type of processing, and practice. Cerebral. Cortex 7, 374 – 385. Gutierrez, S., Corsi-Cabrera, M., 1988. EEG activity during performance of cognitive tasks demanding verbal and/or spatial processing. Int. J. Neurosci. 42, 149 – 155. Harmony, T., Ferna´ndez, T., Silva, J., Bernal, J., Dı´az-Comas, L., Reyes, A., Marosi, E., Rodrı´guez, M., 1996. EEG delta activity: An indicator of attention to internal processing during performance of mental tasks. Int. J. Psychophysiol. 24, 161 – 171. Harmony, T., Ferna´ndez, T., Silva, J., Bosch, J., Valde´s, P., Ferna´ndez-Bouzas, A., Gala´n, L., Aubert, E., Rodrı´guez, D., 1999. Do specific EEG frequencies indicate different processes during mental calculation? Neurosci. Lett. 266, 25 – 28. Harmony, T., Ferna´ndez, T., Ferna´ndez-Bouzas, A., Silva-Pereyra, J., Bosch, J., Dı´az-Comas, L., Gala´n, L., 2001. EEG changes during word and figure categorization. Clin. Neurophysiol. 112, 1486 – 1498. Hitch, G.J., 1978. The role of short-term working memory in mental arithmetic. Cogn. Psychol. 10, 302 – 323. Jonides, J., Smith, E.E., 1997. The architecture of working memory. In: Rugg, M.D. (Ed.), Cognitive Neurosciences. MIT Press, Cambridge, pp. 243 – 276. Kakizaki, T., 1984. Relationship between EEG amplitude and subjective rating of task strain during performance of a calculating task. Eur. J. Appl. Physiol. 53, 206 – 212. Klimesch, W., 1999. EEG alpha and theta oscillations reflect cognitive and memory performance: a review and analysis. Brain Res. Rev. 29, 169 – 195. Legewie, H., Simonova, O., Creutzfeldt, O.D., 1969. EEG changes performance of various tasks under open- and closed-eye conditions. Electroencephalogr. Clin. Neurophysiol. 27, 470 – 479. Luria, A.R., 1978. El Cerebro en Accio´n Pueblo y Educacio´n, La Habana, Cuba. Mazziotta, J.C., Toga, A., Evans, A.C., Fox, P., Lancaster, J., 1995. A probabilistic atlas of the human brain: theory and rationale for its development. NeuroImage 2, 89 – 101. McCarthy, R.A, Warrington, E.K., 1990. Cognitive Neuropsychology: A Clinical Introduction. Academic Press, San Diego, CA. McCloskey, M., 1992. Cognitive mechanisms in numerical processing: evidence from acquired Dyscalculia. Cognition 44, 107 – 157. Mesulam, M.M., 1990. Large-scale neurocognitive networks and distributed processing for attention, language, and memory. Ann. Neurol. 28, 597 – 613. Mulert, C., Gallinat, J., Dorn, H., Herrmann, W.R., Winterer, G., 2003. The relationship between reaction time, error rate and
216
T. Harmony et al. / International Journal of Psychophysiology 53 (2004) 207–216
anterior cingulate cortex activity. Int. J. Psychophysiol. 47, 175 – 183. Pascual-Marquı´, R.D., 1994. Low resolution brain electromagnetic tomography. Brain Topogr. 7, 180 (Proceedings of the International Seventh Swiss Brain Mapping Meeting, March 1994, Zurich). Pascual-Marquı´, R.D., 1995. LORETA in 3D solution space. ISBET Newslett. (6), 22 – 24. Pascual-Marquı´, R.D., Michel, C.M., Lehmann, D., 1994. Low resolution electromagnetic tomography: a new method for localizing electrical activity in the brain. Int. J. Psychophysiol. 18, 49 – 65. Posner, M.I., 1988. Structures and functions of selective attention. In: Boll, T., Bryant, B. (Eds.), Master Lectures in Clinical Neuropsychology and Brain Function: Research, Measurement, and Practice. American Psychological Association, Washington, DC, pp. 171 – 202. Prichep, L.S., John, E.R., MeeLee, T., 2001. Localization of deep white matter lymphoma using VARETA: a case study. Clin. Electroencephalogr. 32, 62 – 65. Riera, J., Fuentes, M.E., Aubert, E., Dı´az, D., 1997. Solving the forward problem: spherical vs realistic electric lead field. Biomed. Techn. 42, 223 – 226. Sasaki, K., Nambu, A., Tsujimoto, T., Matsuzaki, R., Kyuhou, S., Gemba, H., 1996. Studies on integrative functions of the human frontal association cortex with MEG. Cogn. Brain Res. 5, 165 – 174.
Scherg, M., Von Cramon, D., 1985. Two bilateral sources of the late AEP as identified by a spatio-temporal dipole model. Electroencephalogr. Clin. Neurophysiol. 62, 32 – 44. Sternberg, S., 1966. High speed scanning in human memory. Science 153, 52 – 654. Tucker, D.M., Dawson, S.L., Toth, D.L., Penland, J.G., 1985. Regional changes in EEG, power and coherence during cognition: intensive study of two individuals. Behav. Neurosci. 99, 564 – 577. Valde´s, P., Martı´, F., Garcı´a, F., Casanova, R., 1996. Variable resolution electromagnetic tomography. In: Wood, C. (Ed.), Proceedings of the Tenth International Conference on Biomag´ 96. Pergamon Press, Santa Fe, New Mexico, netism, BIOMAG pp. 137 – 141. Valde´s-Sosa, P., Riera, J., Casanova, R., 1996. Spatio temporal distributed inverse solutions. In: Wood, C. (Ed.), Proceedings of the Tenth International Conference on Biomagnetism, BIO´ 96, Santa Fe, New Mexico. MAG Valentino, D.A., Arruda, J.E., Gold, S.M., 1992. Comparison of EEG and response accuracy in good vs. poorer performers during a vigilance task. Int. J. Psychophysiol. 15, 123 – 133. Worsley, K.J., Marrett, S., Neelin, P., Vandal, A.C., Friston, K.J., Evans, A.C., 1995. A unified statistical approach for determining significant signals in images of cerebral activation. Hum. Brain Mapp. 4, 58 – 73.
