- Email: [email protected]
Electroencephalographic peak alpha frequency correlates of cognitive traits
Neuroscience Letters 371 (2004) 60–63
Electroencephalographic peak alpha frequency correlates of cognitive traits Efthymios Angelakisa,∗ , Joel F. Lubarb , Stamatina Stathopoulouc a
Department of Psychology, Drexel University, MS 626, 245 N. 15th Street, Philadelphia, PA 19102-1192, USA b Department of Psychology, University of Tennessee, Knoxville, TN, USA c Department of Psychology, Chestnut Hill College, Philadelphia, PA, USA Received 2 June 2004; received in revised form 13 July 2004; accepted 17 August 2004
Abstract EEG peak alpha frequency (PAF) has been shown to differentiate groups of adults with higher memory performance from those of lower performance, groups of children with advanced reading ability from matched controls, and to predict state-dependent working memory. The present study attempted to explore PAF as a predicting variable for verbal and attentional cognitive trait abilities in young adults. Nineteen undergraduate students had their EEG recorded during initial rest, reading, and post-reading rest, and at a different day were evaluated on reading, vocabulary, and attentional performance. Results showed significant correlations of reading vocabulary and response control with PAF during reading and post-reading recordings, but not during initial rest. PAF may reflect some general cognitive ability that is not necessarily memory or reading, possibly response control or the ability to acquire vocabulary. It is suggested that cognitive ability traits may reflect the ability to induce cognitive states. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: EEG; Trait; Cognitive ability; Reading; Vocabulary; Response control
EEG peak alpha frequency (PAF) has been shown to differentiate non-clinical controls from groups with various neurological pathologies, including Alzheimer’s dementia [9,10], schizophrenia [5], chronic fatigue syndrome [4], hemispheric stroke [7], and traumatic brain injury [3]. Moreover, PAF has been shown to be higher in neurologically healthy adults with higher memory performance Wechsler Memory Scale (WMS) from those of lower performance [8], and higher in children with advanced reading ability from matched controls [11]. In both the latter cases, higher performance/ability was associated with higher PAF by approximately 1 Hz. PAF has also been shown to reflect individual differences. Klimesch et al. [9] have reported a correlation of PAF recorded at rest with the WMS amongst Alzheimer’s patients, as well as among non-clinical young adults during rest and a memory recognition task, whereas a recent study showed ∗
Corresponding author. Tel.: +1 215 762 3169; fax: +1 215 762 7977. E-mail addresses: [email protected], [email protected] (E. Angelakis). 0304-3940/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2004.08.041
PAF to correlate with working memory (Digit Span) performance across the lifespan (11–70 years) [6]. These studies suggest that there may be some cognitive process common to memory and reading that is reflected in EEG PAF. The present study investigated PAF as a continuous variable for predicting cognitive traits on a number of measures related to memory and reading, including phonetic reading, reading comprehension, reading vocabulary, attention, and response control in a group of non-clinical young adults. Moreover, it was aimed to investigate the recording conditions under which PAF may show such potential. We [3] have recently shown that different recording conditions reflect different correlates of PAF with cognitive preparedness, i.e., the performance potential for higher cognitive functions. Specifically, pre-task resting conditions reflected state differences within non-clinical young adults, and post-task recordings reflected trait differences between individuals with TBI and neurologically intact matched controls. Therefore, the present study attempted to correlate cognitive traits with PAF during pre-task, task, and post-task recordings. To minimize
E. Angelakis et al. / Neuroscience Letters 371 (2004) 60–63
the contribution of state cognitive preparedness on the variance of cognitive performance explained by PAF, psychometric testing was administered at a different day from EEG recording. The present article will provide a summary of the methods employed. For a thorough and detailed description, see the report by Angelakis and Lubar [1]. Nineteen psychology college students were included, 12 males and 7 females, all volunteering for extra credit. A Lexicor Neurosearch 24 EEG recorder (Lexicor Inc., GA) using an electrode cap with 19 electrodes placed according to the 10/20 system. The recording montage was referential to physically linked ears, and the amplifiers were band-pass filtered at 0.5–32 Hz. EEG was recorded at 128 samples per second for the initial eyes-closed resting baseline with a band-pass filter set at 0.5–32, and at 256 samples per second for all other recordings, with a bandpass filter set at 0.5–64 Hz. Reading materials for the three experimental tasks were developed in our laboratory. Three pieces from Homer’s Odyssey translated in English were used to selectively engage participants into visual, phonological, and semantic processing. Participants were asked to identify target words following different rules for each processing modality. Visual reading required the identification of four-letter words that include at least one “a” (e.g., have); phonological reading required the identification of words that included the sound “k” (as in cross or peak); and semantic reading required the identification of nouns that refer to a non-animate material object or entity (e.g., table or ocean). Texts were selected so that they were narrative, easy to read, and with a minimum number of names. Moreover, all three texts contained 20 (±1) target words for all three reading requirements, but in randomly different positions. Reading materials were presented with a Pentium computer with a 17 in. color screen. Psychometric tests included four subtests from the Woodcock–Johnson Achievement Battery Revised (WJ), namely the Word Attack (WA) subtest for the assessment of phonic, structural and auditory processing skills, the LetterWord Identification (LWI) subtest for the assessment of pronunciation and paralexic reading, the Passage Comprehension (PC) subtest for the assessment of reading comprehension skills, and the Reading Vocabulary (RV) subtest for the assessment of word semantic/conceptual skills. Participants were also evaluated for their vocabulary with the Vocabulary (V) subtest of the Wechsler Adult Intelligence Scale Revised (WAIS). Finally, participants were given the Intermediate Visual & Auditory continuous performance test (IVA), including the Attention (ATT) index for the assessment of vigilance, focus, and speed; and the Response (RES) index for the assessment of prudence, consistency, and stamina (BrainTrain, Inc.). Participation was completed in two sessions on different days. All EEG recordings were completed within the first day of participation, whereas psychometric evaluation was administered on a following day, no more than 1 week post-EEG. During the first day, 19-channel EEG activity was
61
recorded in the following order: first, during an eyes-closed resting condition (ECB); second, during an eyes-open resting condition (EOB), where participants were instructed to focus on the notepad window on the computer screen, while no text was running; then, five reading tasks (the three Odyssey texts, a list of misspelled words and a list of numbers) and a post-task eyes-open resting condition (PTR) were administered in a counterbalanced order across participants. To avoid confounding of order effects and text related effects, the three Odyssey texts were always presented in the same order, but for different reading requirements (i.e., visual, phonological, or semantic), in a counterbalanced order. Each recording lasted 3.3 min. Before recording each reading task, a practice task was administered for 30 s. All reading was silent. While reading, participants were responding to target word identification by pressing a key on the computer keyboard with their right hand. The procedure was completed within 120 min. Digital EEG was processed by Fast-Fourier Transformation (FFT) with cosine tapering (Hanning window). Raw EEG data were visually inspected and all epochs including artifacts were removed from further analysis. Average PAF was computed for all artifact-free epochs (regardless of whether they contained marked responses or not) for each 3.3 min recording, using the Workstation EEG analysis software (NovatechEEG, Inc.). PAF was defined as the discrete frequency that had the highest average amplitude in each recording, within the range of 7–13 Hz for each of the 19 channels. In order to introduce more variability, all 19 channels for each recording were averaged to give a single PAF score, which created a much more continuous range of values. Previous studies with PAF have shown their results widely distributed at most scalp locations [9,11]. Psychometric performance was then correlated with PAF from each recording condition separately. Since not all data were normally distributed, Spearman’s “rho” non-parametric test was computed to assess statistical significance. Performance on all reading tasks was found to be above 50% correct for all participants (i.e., at least 10 out of 20 targets were identified) and for the purposes of the present analysis we considered this success rate as acceptable (this is a much higher than chance success, since target words were very few compared to non-target words). However, semantic reading produced significantly less omission and more commission errors compared with visual and phonetic reading. Visual and phonetic tasks did not differ significantly in either omission or commission errors. Table 1 summarizes the results. Only phonetic reading and PTR showed significant correlations of PAF with psychometric performance. Specifically, PTR was correlated with IVARES (r = 0.868, p < 0.001), with WAIS-V (r = 0.645, p < 0.01), with WJ-RV (r = 0.559, p < 0.05), with IVA-ATT (r = 0.521, p < 0.05), and with WJ-PC (r = 0.471, p < 0.05). Phonetic reading was correlated with WJ-RV (r = 0.723, p < 0.001), with IVA-RES (r = 0.557, p < 0.01), and with WAISV (r = 0.403, p < 0.05). From these correlations, only PTR with IVA-RES and PHO with WJ-RV remained statistically
62
E. Angelakis et al. / Neuroscience Letters 371 (2004) 60–63
Table 1 Correlations of PAF with psychometrics
WJ-WA WJ-LWI WJ-PC WJ-RV WAIS-V IVA-ATT IVA-RES ∗ ∗∗ ∗∗∗
ECB
EOB
VIS
PHO
SEM
PTR
ns ns ns ns ns ns ns
ns ns ns ns ns ns ns
ns ns ns ns ns ns ns
ns ns ns r = 0.723∗∗∗ r = 0.403∗ ns r = 0.557∗∗
ns ns ns ns ns ns ns
ns ns r = 0.471∗ r = 0.559∗ r = 0.645∗∗ r = 0.521∗ r = 0.868∗∗∗
p < 0.05 (significant before correcting a-level for multiple comparisons). p < 0.01 (significant before correcting a-level for multiple comparisons). p < 0.001 (significant after correcting a-level for multiple comparisons).
significant after adjusting the alpha level for multiple comparisons to avoid type-I error using the Bonferonni method (0.05:42 = 0.001) (Fig. 1). The present results show that PAF has the potential to predict individual trait differences in reading vocabulary, response control, and possibly reading comprehension, and attention. Even though the latter two abilities did not pass the corrected significance threshold, they may be true, since the correction for multiple comparisons that was implemented is very conservative, and may have induced type-II error. This is supported further by the fact that all correlations with p < 0.05 are found in the two conditions that yielded significant correlations after Bonferonni correction. For these reasons, the reader is called to consider of interest all correlations below 0.05. The fact that PTR showed the highest correlation, and that it correlated with more psychometrics than phonetic reading implies that alpha frequency EEG during post-task rest may be more powerful to assess cognitive performance than EEG during initial rest or during actual processing. In a different analysis of the same dataset [1], amplitude of high alpha (10–12 Hz) from all three reading conditions was different (lower) from PTR but not from initial rest. Likewise, in the present study conventional initial resting conditions did not show a relationship between PAF and cognitive performance.
On the other hand, Angelakis et al. [3] did find significant correlations of pre-task PAF with working memory, but showed this to be true only for Digit Span administered immediately after EEG recording, thus related to state, but not necessarily trait. Clark et al. [6] also found PAF during rest to correlate with Digit Span, which again may reflect state qualities, even though in their methods they do not clarify the exact timing and sequence of EEG recording and Digit Span administration. Similarly, Klimesch et al. [9] showed significant correlations between PAF during rest (apparently not post-task) with memory performance, but these authors also did not report the time relationship between memory testing and EEG recording. In the present study, however, EEG and psychometrics were recorded at different days, and thus any relationship between the two is more prone to reveal a stable trait of the brain. Still, since in the present study PAF and cognitive performance were correlated during reading and post-reading rest, it is possible that the trait of cognitive performance is the ability to put oneself into the state of cognitive preparedness. This phenomenon seems to be restricted to the alpha frequency, as shown in yet another analysis of the same dataset, where slower rhythms (1–8 Hz) showed an opposite pattern, i.e., they differentiated reading from initial rest but not from PTR [2]. In summary, PAF shows strong relationships with the ability to acquire vocabulary and to control response, which points to some general cognitive ability that is not necessarily memory or reading. It is possible that what is commonly related to PAF in the previous studies is the ability to acquire vocabulary. PAF highest correlation with response control, although not significantly larger than that with vocabulary, poses the question whether response control is related to all three cognitive abilities studied, i.e., vocabulary, memory, and reading performance. Furthermore, PAF correlates with trait cognitive ability during task and post-task rest conditions, but not during pre-task spontaneous rest, suggesting that cognitive traits may reflect the ability to induce cognitive states.
Fig. 1. Scatterplots between (a) peak alpha frequency recorded during post-task rest and response control (left); and (b) peak alpha frequency recorded during phonetic reading and reading vocabulary (right).
E. Angelakis et al. / Neuroscience Letters 371 (2004) 60–63
Acknowledgements We like to thank Philip Vanlandingham and Kerry Towler for assisting with stimuli design and psychometric testing, and NovatechEEG for free supply of EEG analysis software.
References [1] E. Angelakis, J.F. Lubar, Quantitative electroencephalographic amplitude measures in young adults during reading tasks and rest, J. Neurother. 6 (2002) 5–19. [2] E. Angelakis, J.F. Lubar, J. Frederick, S. Stathopoulou, The role of slow-wave electroencephalographic amplitude measures in reading, J. Neurother. 5 (2001) 5–25. [3] E. Angelakis, J.F. Lubar, S. Stathopoulou, J. Kounios, Peak alpha frequency: an electroencephalographic measure of cognitive preparedness, Clin. Neurophysiol. 15 (2004) 887– 897. [4] K.M. Billiot, T.H. Budzynski, F. Andrasik, EEG patterns and chronic fatigue syndrome, J. Neurother. 2 (1997) 20–30.
63
[5] J.M. Canive, J.D. Lewine, J.C. Edgar, J.T. Davis, G.A. Miller, F. Torres, V.B. Tuason, Spontaneous brain magnetic activity in schizophrenia patients treated with aripiprazole, Psychopharmacol. Bull. 34 (1998) 101–105. [6] C.R. Clark, M.D. Veltmeyer, R.J. Hamilton, E. Simms, R. Paul, D. Hermens, E. Gordon, Spontaneous alpha peak frequency predicts working memory performance across the age span, Int. J. Psychophysiol. 53 (2004) 1–9. [7] C. Juhasz, A. Kamondi, I. Szirmai, Spectral, EEG analysis following hemispheric stroke: evidences of transhemispheric diaschisis, Acta Neurol. Scand. 96 (1997) 397–400. [8] W. Klimesch, EEG-alpha rhythms and memory processes, Int. J. Psychophysiol. 26 (1997) 319–340. [9] W. Klimesch, H. Schimke, G. Ladurner, G. Pfurtscheller, Alpha frequency and memory performance, J. Psychophysiol. 4 (1990) 381–390. [10] S. Passero, R. Rocchi, G. Vatti, L. Burgalassi, N. Battistini, Quantitative EEG mapping, regional cerebral blood flow and neuropsychological function in Alzheimer’s disease, Dementia 6 (1995) 148–156. [11] S.M. Suldo, L.A. Olson, J.R. Evans, Quantitative EEG evidence of increased alpha peak frequency in children with precociously reading ability, J. Neurother. 5 (2001) 39–50.
Electroencephalographic peak alpha frequency correlates of cognitive traits Efthymios Angelakisa,∗ , Joel F. Lubarb , Stamatina Stathopoulouc a
Department of Psychology, Drexel University, MS 626, 245 N. 15th Street, Philadelphia, PA 19102-1192, USA b Department of Psychology, University of Tennessee, Knoxville, TN, USA c Department of Psychology, Chestnut Hill College, Philadelphia, PA, USA Received 2 June 2004; received in revised form 13 July 2004; accepted 17 August 2004
Abstract EEG peak alpha frequency (PAF) has been shown to differentiate groups of adults with higher memory performance from those of lower performance, groups of children with advanced reading ability from matched controls, and to predict state-dependent working memory. The present study attempted to explore PAF as a predicting variable for verbal and attentional cognitive trait abilities in young adults. Nineteen undergraduate students had their EEG recorded during initial rest, reading, and post-reading rest, and at a different day were evaluated on reading, vocabulary, and attentional performance. Results showed significant correlations of reading vocabulary and response control with PAF during reading and post-reading recordings, but not during initial rest. PAF may reflect some general cognitive ability that is not necessarily memory or reading, possibly response control or the ability to acquire vocabulary. It is suggested that cognitive ability traits may reflect the ability to induce cognitive states. © 2004 Elsevier Ireland Ltd. All rights reserved. Keywords: EEG; Trait; Cognitive ability; Reading; Vocabulary; Response control
EEG peak alpha frequency (PAF) has been shown to differentiate non-clinical controls from groups with various neurological pathologies, including Alzheimer’s dementia [9,10], schizophrenia [5], chronic fatigue syndrome [4], hemispheric stroke [7], and traumatic brain injury [3]. Moreover, PAF has been shown to be higher in neurologically healthy adults with higher memory performance Wechsler Memory Scale (WMS) from those of lower performance [8], and higher in children with advanced reading ability from matched controls [11]. In both the latter cases, higher performance/ability was associated with higher PAF by approximately 1 Hz. PAF has also been shown to reflect individual differences. Klimesch et al. [9] have reported a correlation of PAF recorded at rest with the WMS amongst Alzheimer’s patients, as well as among non-clinical young adults during rest and a memory recognition task, whereas a recent study showed ∗
Corresponding author. Tel.: +1 215 762 3169; fax: +1 215 762 7977. E-mail addresses: [email protected], [email protected] (E. Angelakis). 0304-3940/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.neulet.2004.08.041
PAF to correlate with working memory (Digit Span) performance across the lifespan (11–70 years) [6]. These studies suggest that there may be some cognitive process common to memory and reading that is reflected in EEG PAF. The present study investigated PAF as a continuous variable for predicting cognitive traits on a number of measures related to memory and reading, including phonetic reading, reading comprehension, reading vocabulary, attention, and response control in a group of non-clinical young adults. Moreover, it was aimed to investigate the recording conditions under which PAF may show such potential. We [3] have recently shown that different recording conditions reflect different correlates of PAF with cognitive preparedness, i.e., the performance potential for higher cognitive functions. Specifically, pre-task resting conditions reflected state differences within non-clinical young adults, and post-task recordings reflected trait differences between individuals with TBI and neurologically intact matched controls. Therefore, the present study attempted to correlate cognitive traits with PAF during pre-task, task, and post-task recordings. To minimize
E. Angelakis et al. / Neuroscience Letters 371 (2004) 60–63
the contribution of state cognitive preparedness on the variance of cognitive performance explained by PAF, psychometric testing was administered at a different day from EEG recording. The present article will provide a summary of the methods employed. For a thorough and detailed description, see the report by Angelakis and Lubar [1]. Nineteen psychology college students were included, 12 males and 7 females, all volunteering for extra credit. A Lexicor Neurosearch 24 EEG recorder (Lexicor Inc., GA) using an electrode cap with 19 electrodes placed according to the 10/20 system. The recording montage was referential to physically linked ears, and the amplifiers were band-pass filtered at 0.5–32 Hz. EEG was recorded at 128 samples per second for the initial eyes-closed resting baseline with a band-pass filter set at 0.5–32, and at 256 samples per second for all other recordings, with a bandpass filter set at 0.5–64 Hz. Reading materials for the three experimental tasks were developed in our laboratory. Three pieces from Homer’s Odyssey translated in English were used to selectively engage participants into visual, phonological, and semantic processing. Participants were asked to identify target words following different rules for each processing modality. Visual reading required the identification of four-letter words that include at least one “a” (e.g., have); phonological reading required the identification of words that included the sound “k” (as in cross or peak); and semantic reading required the identification of nouns that refer to a non-animate material object or entity (e.g., table or ocean). Texts were selected so that they were narrative, easy to read, and with a minimum number of names. Moreover, all three texts contained 20 (±1) target words for all three reading requirements, but in randomly different positions. Reading materials were presented with a Pentium computer with a 17 in. color screen. Psychometric tests included four subtests from the Woodcock–Johnson Achievement Battery Revised (WJ), namely the Word Attack (WA) subtest for the assessment of phonic, structural and auditory processing skills, the LetterWord Identification (LWI) subtest for the assessment of pronunciation and paralexic reading, the Passage Comprehension (PC) subtest for the assessment of reading comprehension skills, and the Reading Vocabulary (RV) subtest for the assessment of word semantic/conceptual skills. Participants were also evaluated for their vocabulary with the Vocabulary (V) subtest of the Wechsler Adult Intelligence Scale Revised (WAIS). Finally, participants were given the Intermediate Visual & Auditory continuous performance test (IVA), including the Attention (ATT) index for the assessment of vigilance, focus, and speed; and the Response (RES) index for the assessment of prudence, consistency, and stamina (BrainTrain, Inc.). Participation was completed in two sessions on different days. All EEG recordings were completed within the first day of participation, whereas psychometric evaluation was administered on a following day, no more than 1 week post-EEG. During the first day, 19-channel EEG activity was
61
recorded in the following order: first, during an eyes-closed resting condition (ECB); second, during an eyes-open resting condition (EOB), where participants were instructed to focus on the notepad window on the computer screen, while no text was running; then, five reading tasks (the three Odyssey texts, a list of misspelled words and a list of numbers) and a post-task eyes-open resting condition (PTR) were administered in a counterbalanced order across participants. To avoid confounding of order effects and text related effects, the three Odyssey texts were always presented in the same order, but for different reading requirements (i.e., visual, phonological, or semantic), in a counterbalanced order. Each recording lasted 3.3 min. Before recording each reading task, a practice task was administered for 30 s. All reading was silent. While reading, participants were responding to target word identification by pressing a key on the computer keyboard with their right hand. The procedure was completed within 120 min. Digital EEG was processed by Fast-Fourier Transformation (FFT) with cosine tapering (Hanning window). Raw EEG data were visually inspected and all epochs including artifacts were removed from further analysis. Average PAF was computed for all artifact-free epochs (regardless of whether they contained marked responses or not) for each 3.3 min recording, using the Workstation EEG analysis software (NovatechEEG, Inc.). PAF was defined as the discrete frequency that had the highest average amplitude in each recording, within the range of 7–13 Hz for each of the 19 channels. In order to introduce more variability, all 19 channels for each recording were averaged to give a single PAF score, which created a much more continuous range of values. Previous studies with PAF have shown their results widely distributed at most scalp locations [9,11]. Psychometric performance was then correlated with PAF from each recording condition separately. Since not all data were normally distributed, Spearman’s “rho” non-parametric test was computed to assess statistical significance. Performance on all reading tasks was found to be above 50% correct for all participants (i.e., at least 10 out of 20 targets were identified) and for the purposes of the present analysis we considered this success rate as acceptable (this is a much higher than chance success, since target words were very few compared to non-target words). However, semantic reading produced significantly less omission and more commission errors compared with visual and phonetic reading. Visual and phonetic tasks did not differ significantly in either omission or commission errors. Table 1 summarizes the results. Only phonetic reading and PTR showed significant correlations of PAF with psychometric performance. Specifically, PTR was correlated with IVARES (r = 0.868, p < 0.001), with WAIS-V (r = 0.645, p < 0.01), with WJ-RV (r = 0.559, p < 0.05), with IVA-ATT (r = 0.521, p < 0.05), and with WJ-PC (r = 0.471, p < 0.05). Phonetic reading was correlated with WJ-RV (r = 0.723, p < 0.001), with IVA-RES (r = 0.557, p < 0.01), and with WAISV (r = 0.403, p < 0.05). From these correlations, only PTR with IVA-RES and PHO with WJ-RV remained statistically
62
E. Angelakis et al. / Neuroscience Letters 371 (2004) 60–63
Table 1 Correlations of PAF with psychometrics
WJ-WA WJ-LWI WJ-PC WJ-RV WAIS-V IVA-ATT IVA-RES ∗ ∗∗ ∗∗∗
ECB
EOB
VIS
PHO
SEM
PTR
ns ns ns ns ns ns ns
ns ns ns ns ns ns ns
ns ns ns ns ns ns ns
ns ns ns r = 0.723∗∗∗ r = 0.403∗ ns r = 0.557∗∗
ns ns ns ns ns ns ns
ns ns r = 0.471∗ r = 0.559∗ r = 0.645∗∗ r = 0.521∗ r = 0.868∗∗∗
p < 0.05 (significant before correcting a-level for multiple comparisons). p < 0.01 (significant before correcting a-level for multiple comparisons). p < 0.001 (significant after correcting a-level for multiple comparisons).
significant after adjusting the alpha level for multiple comparisons to avoid type-I error using the Bonferonni method (0.05:42 = 0.001) (Fig. 1). The present results show that PAF has the potential to predict individual trait differences in reading vocabulary, response control, and possibly reading comprehension, and attention. Even though the latter two abilities did not pass the corrected significance threshold, they may be true, since the correction for multiple comparisons that was implemented is very conservative, and may have induced type-II error. This is supported further by the fact that all correlations with p < 0.05 are found in the two conditions that yielded significant correlations after Bonferonni correction. For these reasons, the reader is called to consider of interest all correlations below 0.05. The fact that PTR showed the highest correlation, and that it correlated with more psychometrics than phonetic reading implies that alpha frequency EEG during post-task rest may be more powerful to assess cognitive performance than EEG during initial rest or during actual processing. In a different analysis of the same dataset [1], amplitude of high alpha (10–12 Hz) from all three reading conditions was different (lower) from PTR but not from initial rest. Likewise, in the present study conventional initial resting conditions did not show a relationship between PAF and cognitive performance.
On the other hand, Angelakis et al. [3] did find significant correlations of pre-task PAF with working memory, but showed this to be true only for Digit Span administered immediately after EEG recording, thus related to state, but not necessarily trait. Clark et al. [6] also found PAF during rest to correlate with Digit Span, which again may reflect state qualities, even though in their methods they do not clarify the exact timing and sequence of EEG recording and Digit Span administration. Similarly, Klimesch et al. [9] showed significant correlations between PAF during rest (apparently not post-task) with memory performance, but these authors also did not report the time relationship between memory testing and EEG recording. In the present study, however, EEG and psychometrics were recorded at different days, and thus any relationship between the two is more prone to reveal a stable trait of the brain. Still, since in the present study PAF and cognitive performance were correlated during reading and post-reading rest, it is possible that the trait of cognitive performance is the ability to put oneself into the state of cognitive preparedness. This phenomenon seems to be restricted to the alpha frequency, as shown in yet another analysis of the same dataset, where slower rhythms (1–8 Hz) showed an opposite pattern, i.e., they differentiated reading from initial rest but not from PTR [2]. In summary, PAF shows strong relationships with the ability to acquire vocabulary and to control response, which points to some general cognitive ability that is not necessarily memory or reading. It is possible that what is commonly related to PAF in the previous studies is the ability to acquire vocabulary. PAF highest correlation with response control, although not significantly larger than that with vocabulary, poses the question whether response control is related to all three cognitive abilities studied, i.e., vocabulary, memory, and reading performance. Furthermore, PAF correlates with trait cognitive ability during task and post-task rest conditions, but not during pre-task spontaneous rest, suggesting that cognitive traits may reflect the ability to induce cognitive states.
Fig. 1. Scatterplots between (a) peak alpha frequency recorded during post-task rest and response control (left); and (b) peak alpha frequency recorded during phonetic reading and reading vocabulary (right).
E. Angelakis et al. / Neuroscience Letters 371 (2004) 60–63
Acknowledgements We like to thank Philip Vanlandingham and Kerry Towler for assisting with stimuli design and psychometric testing, and NovatechEEG for free supply of EEG analysis software.
References [1] E. Angelakis, J.F. Lubar, Quantitative electroencephalographic amplitude measures in young adults during reading tasks and rest, J. Neurother. 6 (2002) 5–19. [2] E. Angelakis, J.F. Lubar, J. Frederick, S. Stathopoulou, The role of slow-wave electroencephalographic amplitude measures in reading, J. Neurother. 5 (2001) 5–25. [3] E. Angelakis, J.F. Lubar, S. Stathopoulou, J. Kounios, Peak alpha frequency: an electroencephalographic measure of cognitive preparedness, Clin. Neurophysiol. 15 (2004) 887– 897. [4] K.M. Billiot, T.H. Budzynski, F. Andrasik, EEG patterns and chronic fatigue syndrome, J. Neurother. 2 (1997) 20–30.
63
[5] J.M. Canive, J.D. Lewine, J.C. Edgar, J.T. Davis, G.A. Miller, F. Torres, V.B. Tuason, Spontaneous brain magnetic activity in schizophrenia patients treated with aripiprazole, Psychopharmacol. Bull. 34 (1998) 101–105. [6] C.R. Clark, M.D. Veltmeyer, R.J. Hamilton, E. Simms, R. Paul, D. Hermens, E. Gordon, Spontaneous alpha peak frequency predicts working memory performance across the age span, Int. J. Psychophysiol. 53 (2004) 1–9. [7] C. Juhasz, A. Kamondi, I. Szirmai, Spectral, EEG analysis following hemispheric stroke: evidences of transhemispheric diaschisis, Acta Neurol. Scand. 96 (1997) 397–400. [8] W. Klimesch, EEG-alpha rhythms and memory processes, Int. J. Psychophysiol. 26 (1997) 319–340. [9] W. Klimesch, H. Schimke, G. Ladurner, G. Pfurtscheller, Alpha frequency and memory performance, J. Psychophysiol. 4 (1990) 381–390. [10] S. Passero, R. Rocchi, G. Vatti, L. Burgalassi, N. Battistini, Quantitative EEG mapping, regional cerebral blood flow and neuropsychological function in Alzheimer’s disease, Dementia 6 (1995) 148–156. [11] S.M. Suldo, L.A. Olson, J.R. Evans, Quantitative EEG evidence of increased alpha peak frequency in children with precociously reading ability, J. Neurother. 5 (2001) 39–50.