2. 2 K. Yun et al. / Psychiatry Research: Neuroimaging xxx (2012) xxx–xxx
performance (Newton et al., 2004). Power spectrum analysis revealed (N = 20, average age = 34.5± 7.7 years, range= 23–48 years, all males)
an apparent EEG slowing in MA abusers (Newton et al., 2003, 2004), but were recruited from Bugok National Hospital in South Korea. The MA
correlations with abuse patterns and social factors were not examined. abusers were hospitalized patients who met the Diagnostic and
While pre-clinical and clinical investigations have shown that Statistical Manual of Mental Disorders, 4th edition (DSM-IV) (Spitzer
methamphetamine (MA) causes long-term damage to the DA reward et al., 1994) criteria for lifetime MA abuse (N = 31) or dependence
circuits resulting in motor and cognitive deficits (Volkow et al., 2001b; (N = 17). There were no significant correlations between abused and
Johanson et al., 2006; McCann et al., 2008), but little is known about dependent patients in the period of MA use (p = 0.970), the age of first
dynamical disturbance of cortical network in MA abusers. The aim of the abuse (p = 0.437), and the cumulative amount of MA (p = 0.820). All
present study was to determine whether abstinent MA abusers exhibit subjects were men only because of extensive prior evidence that MA-
alterations in complexity of the EEG. Tononi et al.(1998) suggested that induced neurotoxicity is sex-specific, particularly greater in men than in
optimal brain functioning requires the dynamic interplay between local women (Wagner et al., 1980; Dluzen et al., 2002; D'Astous et al., 2005;
specialization and global integration of brain activity. They proposed Kim et al., 2005b). They showed no signs of neurological abnormalities
that this optimal state produces complex activity and that a neural such as seizure, dyskinesia, or coma. MA abusers were excluded if they
complexity measure is capable of estimating the optimal balance had a past or present history of a comorbid psychiatric illness except
between localization and integration of neural networks (Tononi et al., substance abuse (DSM-IV axis I or II diagnosis). After complete
1998; Tononi and Edelman, 1998; Sporns et al., 2000). Indeed, reduced description of the study, written informed consent was obtained from
complexity of EEG patterns has been reported in patients with all subjects prior to participation, in compliance with the procedures of
Alzheimer's disease (Jeong et al., 1998; Abasolo et al., 2005), the Ethics Committee of the Bugok National Hospital.
schizophrenia (Roschke et al., 1994; Breakspear et al., 2003; Paulus For the accuracy of patient profiles, detailed clinical information on
and Braff, 2003; Keshavan et al., 2004; Micheloyannis et al., 2006), and MA use patterns was obtained through interviews with the abuser and
depression (Roschke et al., 1994; Thomasson et al., 2002; Bob et al., his family members as well as by referral to patient medical records
2006; Fingelkurts et al., 2007), many of whom are hypothesized to suffer using the Structured Clinical Interview for DSM-IV (SCID). The clinical
from reduced functional connectivity between cortical regions. A information includes the period of MA use, the age of first abuse, and
previous study also found that EEG complexity is reduced as sleep other substance abuse such as nalbuphine, nicotine, alcohol, and
goes deeper and increased during REM sleep (Burioka et al., 2005b). inhalant solvents. A heavy drinker was defined as consuming on
Thus, in this study, we examined the interplay between the functional average more than four drinks daily, and a heavy smoker was defined as
integration and segregation of cortical networks and consequently the smoking two or more packs per day. We also obtained the clinical
efficiency of information processing the cortex through quantification of information about sexual exploitation and criminal records. In the MA
the complexity in EEG patterns in MA abusers. binge abuse cycle, during the initial response of the rush, the MA
To estimate complexity of EEG patterns in MA abusers, we used abuser's heartbeat races and metabolism, blood pressure, and pulse
Approximate Entropy (ApEn), an information-theoretic measure of increase. During this "high," they partake in a wide range of risky
irregularity. ApEn can stably quantify the complexity of a noisy and behaviors that include having an increased level of sexual behavior
short time series as in physiological recordings (Pincus, 1991; Pincus, (sexual exploitation) and that incorporate unsafe behaviors, driving, or
1995). Several studies have reported that ApEn can be used to discern committing crimes.
various neuropsychiatric conditions such as Alzheimer's disease, These evaluations were performed within 3 days of the EEG
coma (Abasolo et al., 2005; Lin et al., 2005), and epilepsy examination by a trained research psychiatrist blind to the EEG data.
(Radhakrishnan and Gangadhar, 1998; Hornero et al., 1999; Burioka All MA abusers had taken MA intravenously for at least least 2 years,
et al., 2005a). A previous study also reported that ApEn analysis of and each subject had been abstinent for more than 6 days at the time
heart rate in cocaine abusers showed reduced complexity, suggesting of the EEG examination. The drug was only to be taken intravenously
that impaired function, isolation and network diminution are since this is the most common method of use in South Korea (Chung
manifest across multiple axes (Newlin et al., 2000). et al., 2004). Urine tests were performed just prior to EEG recording
MA is known to induce a variety of symptomatic behaviors during for the proof of negative current intoxication.
intoxication or withdrawal that include irritability, anxiety, excitement, The controls were recruited from the local community in Bugok
hallucinations, paranoia (both delusional and psychotic), and aggressive and Busan, South Korea. They were group-matched with the MA
behavior. Social abnormalities such as criminal misconduct or sexual abusers by age, sex, and socioeconomic status (income: b$25,000;
intercourse are also observed in MA abusers during MA intoxication. education: college or high school graduates). They had no history of
Psychotic behaviors of MA users are correlated with their ages, possibly MA use or abuse of other substances. None had a personal or familial
associated with disturbance of neurotransmitters in cortical–subcortical history of psychiatric illness based on an unstructured interview
circuits during the aging process (Chen et al., 2003). Another conducted by a trained psychiatrist. No subjects could be taking
commonality found among MA abusers is the consumption of other medication at the time of the study, and all had negative urine toxin
substances such as alcohol and nicotine during MA intoxication. These screens to ensure the absence of psychoactive drug use. Controls were
multifaceted factors result in the difficulties in treating MA dependen- excluded if they reported drinking alcohol more than once per day or
cies in abusers. Despite the prominent and detrimental effects of MA on N40 g per week or reported smoking more than 10 cigarettes per day.
the nervous systems and social behavior, few neuroimaging or After evaluation of the EEG, IQ was measured for all subjects as the
electrophysiological studies have been performed to investigate the psychometric index of intellectual functioning (Sattler, 2001) using
relationship between cortical alterations and critical factors including the Korean–Wechsler Adult Intelligence Scale (K-WAIS)-Short form
abuse patterns and social behaviors. Therefore, we aimed to determine (Yeom et al., 1992). The four sections of this shortened test required
the association between the complexity of EEG patterns in MA abusers the subjects to answer questions about the information that was
and their drug abuse patterns. provided, picture completion, arithmetic, and block designs. It was
revised and given in Korean. The estimated IQ score was normalized
2. Methods according to the respondent's age.
2.1. Subjects 2.2. EEG recording
Currently abstinent MA abusers (N = 48, average age = 36.7 ± EEGs were recorded from 16 channels using EEG amplifier (Model 9
5.8 years; range= 26–49 years, all males) and 20 control subjects EEG Grass Instrument Co.) in the morning (10:30–11:30 AM) to
Please cite this article as: Yun, K., et al., Decreased cortical complexity in methamphetamine abusers, Psychiatry Research: Neuroimaging
(2012), doi:10.1016/j.pscychresns.2011.07.009
3. K. Yun et al. / Psychiatry Research: Neuroimaging xxx (2012) xxx–xxx 3
minimize the circadian and homeostatic modulation of wakefulness Table 1
(Taillard et al., 2003). Subjects were instructed to lie with their eyes Demographic profile of the study samples.
closed, think of nothing in particular, and not fall asleep in a sound- MA group (n = 48) Control group (n = 20)
attenuated room. Ag-AgCl electrodes were placed according to the
Age (years) 37.0 (5.8) 34.5 (7.7)
international 10–20 System. Data were collected using a Cz referential IQ (points) 97.8 (6.1) 116.5 (5.3)⁎
montage and were digitized at 200 EEG samplings per second which Duration of MA use (years) 11.8 (6.5) N/A
were obtained by the QUINGY, MASS system (bandpass filtered 0.3– Abstinence period (days) 30.5 (27.2) N/A
Total amount in a previous year (g) 1.125 (1.095) N/A
70 Hz). At least 20 min of EEG activity was recorded and a minimum of
30 s of artifact-free EEG was selected by visual inspection and then was Means are presented with standard deviations in parentheses.
⁎ p b 0.05.
analyzed. To obtain maximally long stationary EEG data, the first 2 min
of data were discarded. Three epochs of 10-s of EEG recordings were
randomly selected to estimate the ApEn values and their means were between the groups. Thus, we employed ‘Levene's Test for Equality of
used. An additional three epochs of 10 s of EEG recordings were Variances’ to select a proper statistic. Levene's Test with p b 0.05
randomly selected. We found that the ApEn values were robust to epoch indicated that variance was not homogeneous. If Levene's Test indicated
selection and the results were consistent in that MA abusers exhibited that variances were homogeneous between groups, then Student's t-
decreased ApEn values compared with those of the healthy subjects in test was used. Otherwise, Welch's t-test was used. Welch's t-test is a
all 15 channels (p b 0.001) except F4 (p N 0.05). There was no effect of variation of Student's t-test intended for use with two samples having
epoch selection (p N 0.3). Using stationary EEG data and random epoch unequal variances.
selection might help to control the possibly variable cognitive processes
involved in the resting state. Moreover, robustness of the ApEn results to 3. Results
epoch selection might indicate that a rather consistent cognitive state
was maintained in our resting supine condition. 3.1. Decreased cortical complexity of MA abusers
2.3. Approximate Entropy analysis Table 1 summarizes the demographic characteristics displaying the
similarities between the former MA abuser and control groups. The two
The recorded EEG data were reformatted offline to compute the groups did not differ significantly in age, but there was a difference in
power spectrum (see supplementary material for power spectrum their IQ levels (p b 0.05). To assess the possible presence of abnormal
analysis and results) and ApEn values. First introduced by Pincus, the complexity of cortical activity in MA abusers, we estimated the ApEn
ApEn is an index that quantifies the irregularity or complexity of a values of EEG patterns in 48 MA abusers and compared them with those
dynamical system (Pincus, 1991). It is particularly efficient to use with of 20 control subjects. We found that MA abusers exhibited decreased
short and noisy time-series data such as physiological data. The ApEn ApEn values compared with those of the healthy subjects in all 16
measures the logarithm frequency with which neighborhoods of channels (Student's t-test; p b 0.00001) (Fig. 1). We also calculated a
temporal patterns of length m that are within a certain distance (r) in multivariate general linear model (dependent variables: ApEn values of
phase space remain close together (br) for patterns that are augmented 16 channels, fixed factor: MA abusers/controls, covariate: IQ, smoking
by one point of time (i.e., for patterns of length m + 1) (see amount). Including IQ and smoking amount as a covariate decreased the
supplementary material for detailed algorithm). Thus, smaller values statistical difference between MA and control groups. However, most
of the ApEn imply stronger regularity or persistence in a time series. channels were still significantly different between groups (Fp1, T6, O1,
Conversely, larger values of the ApEn signify the presence of greater O2, p b 0.0001; T5, p b 0.001; Fp2, P3, P4, F7, F8, T3, p b 0.01; F3, C3,C4, T4,
fluctuations, or irregularity, in a time series. ApEn values of EEG are p b 0.05).
possibly determined by the balance between the functional segrega- To investigate the effects of severity of MA abuse on the ApEn values
tions and integrations of cortical regions. of EEGs, we classified the MA group into two separate subgroups based
on the period of MA abuse and recent intake amount: a moderate and
2.4. Statistical analysis a severe MA abuse group. The “severe MA abusers” (average age was
37.7± 5.52 years; age range: 29–49 years; the average cumulative
Group comparisons of the demographic variables were conducted amount of MA was 14.3 ± 15.2 g) consisted of subjects who used MA
using t-tests. The average values of the ApEn of the EEG in the MA- for at least 6 years or more than 0.75 g of MA injection in the most recent
dependent group and the control group are presented as “mean ± year, compared with ‘moderate MA abusers’ (average age 30.6 ±
standard error” across all subjects. The one-way analysis of variance 3.31 years; age range 26–35 years; average MA cumulative amount
(ANOVA) procedure was used to compare the severe and moderate MA 0.92 ± 0.52 g). ANOVA analysis revealed that severe and moderate MA
abusers and the control group. If the results from the ANOVA achieved abuser groups and the control group exhibited significant differences in
statistical significance (pb 0.05), multiple comparisons were performed ApEn values in all channels (p b 0.0001), indicating the significant
afterwards (LSD test). The effects of abuse patterns, clinical/social
measures, and comorbidity on the cortical complexities of MA abusers
were evaluated separately using t-tests. Pearson's correlation coefficient
was used for the linear correlation analysis, while a statistical software
package (SPSS 11.0.1, SPSS Inc., Chicago, IL, USA) was used. Statistical
significance was defined to have an alpha level of 0.05. Adjacent
electrodes were grouped for pair-wise correction for multiple compar-
isons (Fp1–F3, Fp2–F4, F7–T3, F8–T4, C3–P3, C4–P4, T5–O1, T6–O2).
For normality test of the data, we used the ‘Kolmogorov–Smirnov
test’ before ApEn analysis and did not use outliers to maintain the
normality of the data. Outliers were detected through visual examina-
tion of the scattergrams and normal probability distribution plots. Fewer
than three channels out of the total electrodes for each subject were Fig. 1. Topographic map of average ApEn values in (a) healthy subjects and (b) MA
removed from the study through the normality test. In comparisons of abusers. MA abusers exhibited reduced ApEn values compared with healthy subjects in
two samples (t-test), the variances of some EEG data are not equal all channels (Student's t-test; p b 0.00001).
Please cite this article as: Yun, K., et al., Decreased cortical complexity in methamphetamine abusers, Psychiatry Research: Neuroimaging
(2012), doi:10.1016/j.pscychresns.2011.07.009
4. 4 K. Yun et al. / Psychiatry Research: Neuroimaging xxx (2012) xxx–xxx
Table 2 3.2. Correlations of ApEn values with clinical/social measures
Significant differences of ApEn values of EEGs in MA abusers and controls. Group 1
corresponds to a group having lower ApEn values than those of group 2.
To investigate the associations between ApEn values and clinical
Group 1 Group 2 factors, we divided MA abusers into several groups according to their
MA⁎ Control Fp1, Fp2, F3, C3, P3, P4, F7, F8, T3, T4, sexual histories, their drug-related criminal records, and if they
T5, T6, O1 experienced common symptoms of psychosis (such as hallucinations
Severe-MA⁎ Moderate-MA F3, C3, P3, F7, T5, O1 or delusions). We found no significant correlation between the ApEn
No psychosis⁎ Psychosis Fp1, Fp2, F3, F4, C4, P4, F7, F8, T3, T4,
value of the EEG and the age at which a subject first used MA. The MA
T5, T6, O1, O2
No sex acts⁎ Sex acts C3, P3 patients exhibiting delusions or hallucinations (N = 37) had higher
Drug-related No criminal record Fp2, F4, C3, C4, P3, P4, F7, F8, T3, T4, ApEn values in most channels than the abusers without these
Criminal records⁎ T5, T6, O1, O2 symptoms (Fp1, Fp2, F3, F4, C4, P4, F7, F8, T3, T4, T5, T6, O1, O2;
Heavy smoking⁎ Light smoking C4, P4, F8, T4, T5, T6, O1, O2 p b 0.05) (Fig. 3(a)). MA patients who participated in sexual
No nalbuphine⁎ Nalbuphine Fp2, F4, C3, C4, P3, P4, F7, F8, T3, T4,
intercourse during their MA binges (N = 25) had higher ApEn values
T6, O2
in the centro-parietal areas than the other abusers who did other
Severe-MA: 6 or more years of MA abuse or taking at least 0.75 g of MA in the previous
activities (such as driving cars or playing video games) (C3, P3;
year.
Moderate-MA: under 6 years of MA abuse and taking less than 0.75 g of MA in the p b 0.05) (Fig. 3(b)). The MA patients with drug-related criminal
previous year. records (N = 21) had lower ApEn values than the other subjects (all
Heavy smoking: More than or equal to two packs/day. channels, except Fp1 and F3; p b 0.05) (Fig. S2). However, the MA
Light smoking: Less than two packs/day. patients with other types of criminal records had lower ApEn values
⁎ p b 0.05.
that were limited to the centro-parietal areas (C4, P4; p b 0.05). These
results demonstrate that reductions in cortical complexity (measured
by the ApEn values of the EEGs) of MA abusers are associated with the
influence of the duration of MA abuse and the dosage received during the pathological behaviors that occur during MA abuse.
previous year. To control the possible age difference between groups, we
applied an analysis of covariace (ANCOVA). We selected age, IQ, and
smoking amount as covariates. All channels except F4 were still
significantly different between severe and moderate MA abuser and
control groups (Fp1, T6, O1, O2, p b 0.0001; P4, T3, T5, p b 0.001; F3, C3,
P3, F7, F8, p b 0.01; Fp2, C4, T4, p b 0.05). In post-hoc analyses, the
significant differences were found between the moderate and the severe
MA groups in the left cortical areas (LSD test; F3, t(43.999) = 51.638,
p = 0.042; C3, t(44.421) = 65.289, p = 0.031; P3, t(44.793) = 40.909,
p = 0.049; F7, t(42.743) = 30.308, p = 0.043; T5, t(41.146) = 25.373,
p = 0.037; O1, t(44.979) = 23.934, p = 0.030). The severe MA abuse
group had significantly lower ApEn values than the control group in
most channels (Fp1, Fp2, C3, P3, F7, F8, T3, T5, T6, O1, O2, p b 0.005; F3, C4,
P4, T4, p b 0.05). The moderate MA group exhibited significantly
decreased ApEn values of the EEG in Fp1 channel compared with the
values of the control subjects (LSD test; t(43.592) = 24.366, p b 0.0001)
(Table 2). These results indicate the period-dependent, MA-induced
reduction in cortical complexity in the MA group (Fig. 2).
Fig. 2. ApEn values of the EEG in control, moderate MA abuse, and severe MA abuse groups. Fig. 3. ApEn values of the EEG correlated with (a) the presence of psychoses and (b) sexual
Significant differences were found in cortical regions marked with an asterisk (*) between intercourse during MA intoxication. The asterisk (*) indicates significant differences in
the moderate and the severe MA user groups (F3, C3, P3, F7, T5, O1; p b 0.05). ApEn values (p b 0.05).
Please cite this article as: Yun, K., et al., Decreased cortical complexity in methamphetamine abusers, Psychiatry Research: Neuroimaging
(2012), doi:10.1016/j.pscychresns.2011.07.009
5. K. Yun et al. / Psychiatry Research: Neuroimaging xxx (2012) xxx–xxx 5
3.3. Correlation of the ApEn values with comorbidity ApEn reflects the dynamic balance between the functional integra-
tion and segregation of neural networks (Pincus and Goldberger, 1994;
We determined the relationship between the ApEn values of the Tononi et al., 1994). ApEn appears to decrease when the functional
EEGs in MA abusers and the presence of comorbidity of other cortical network is isolated or decoupled, and this reduced integration
substance abuse. Although other co-morbid abuse of inhalant solvents leads to hypoactivity, or slow signal transmission in the cortical
(N = 7) did not have an effect on the ApEn values, we found heavy networks (Pincus, 1994). This reduction in cortical complexity,
smoking (N = 5) (more than or equal to two packs/day) decreased the particularly in the temporal and frontal regions, probably stems from
ApEn values in the right frontal and central areas and both sides of the the anatomical and functional disconnections among the regions that
temporal and occipital areas (C4, P4, F8, T4, T5, T6, O1, O2, p b 0.05) are most affected by the long-term toxicity of MA. This speculation may
(Fig. S3). The ApEn values revealed that MA abusers who were heavy be consistent with the decreases in the metabolism rates of frontal white
smokers exhibited greater cortical dysfunctions than those who matter in MA abusers, and it suggests that persisting deficiencies can be
smoked fewer cigarettes or did not smoke at all. The abuse of found in the frontal lobes (Kim et al., 2005b). The frontostriatal circuit
nalbuphine hydrochloride or other similar opioid medications (N = 7) might be affected in MA abusers so the decreased striatal metabolism
increased the ApEn values in the global cortical areas (Fp2, F4, C3, C4, rates cause the lower levels of metabolism in the frontal cortical regions.
P3, P4, F7, F8, T3, T4, T6, O2; p b 0.05) (Fig. 4). This hypo-metabolism found in cortical regions possibly decreases the
integrations of the neural networks. This in turn leads to the isolation or
4. Discussion decoupling of the network.
Alternatively, the reductions in dynamical complexity of EEG
In the present study, we detected significantly reduced ApEn values patterns may reflect the toxicity of MA to several cortical monoam-
of EEG patterns (i.e. cortical complexity) in former MA abusers in the inergic neurotransmitter systems such as dopamine and noradrenalin,
global cortical regions compared with those of healthy subjects. Within and thus reduced local activities of cortical regions. The most
the MA abuser group, severe MA abusers had more disproportionate prominent and toxic effects are on the neurites of dopaminergic
reductions in cortical complexity compared with moderate MA abusers. neurons in the deep brain including the striatum, caudate, and
Although the EEG has a poor spatial resolution and possibly topograph- putamen (Kraemer and Maurer, 2002; Kita et al., 2003). Brain damage
ical implications of these findings must be interpreted with extreme by MA is not limited to deep brain structures, but also extends to
caution, this reduction was more prominent in fronto-temporal and global cortical areas (Ciraulo et al., 2003; Kim et al., 2005a). The
occipital regions. MA patients that suffered from delusions and involvement of dopamine in the process of drug addiction is likely to
hallucinations or those that participated in sexual intercourse during be accompanied by functional and structural changes to the circuits,
MA intoxication exhibited increased levels of cortical complexity including the cortical areas. The vast damage of cortical and
compared with those MA abusers that had not experienced similar subcortical regions to which these monoaminergic systems project
psychotic symptoms or sexual intercourse during MA intoxication. MA as found in previous studies may account for the reductions in global
patients having histories of drug-related criminal activities had the most dynamical complexity observed in the former MA abusers of our
decreased cortical complexity among the MA abusers. The MA abusers study. Regardless of the resulting reductions in dynamical complexity
who smoked heavily had decreases in cortical complexity in the right that might be consequences of MA abuse on a functionally coupled
frontal, central, temporal, and occipital areas. The subjects concurrently neural system or the drug's comparable effects on multiple mono-
using nalbuphine hydrochloride had increases in cortical complexity in amine systems, the estimated duration and amount of MA abuse
most areas among the other MA users. These findings indicate overall correlated strongly with global and dynamical EEG complexity across
reductions in dynamical complexity in the cortical networks of MA all portions of these monoaminergic and functional networks. These
abusers, possibly due to the severe and long-lasting presence of toxins in findings suggest that the duration and amount of exposure to MA
the monoaminergic neurotransmitter systems of the brain in MA accounts for a substantial amount of variance found in the severity of
abusers. To the best of our knowledge, this is the first investigation to its long-term and system-wide neurotoxic effects in people.
assess the “severity-dependent dynamical complexity” of EEG patterns In previous reports, delusions and hallucinations found in MA
in former MA abusers and their associations with the subjects' abuse abusers were associated with abnormalities in the dopamine receptors
patterns and other clinical measures. or transporters of the caudate, putamen, and nucleus accumbens (Wada
and Fukui, 1990; Sekine et al., 2001). The ApEn revealed that MA abusers
exhibiting delusions or hallucinations had more complex cortical
activity than the abusers without these psychotic symptoms. It seems
that psychotic symptoms might not only explain certain dopaminergic
system deficits, but the presence of these symptoms might also be the
reason for the reorganization or partial recoveries of the deficits
(Lautenschlager and Forstl, 2001; Kato et al., 2006) resulting from the
toxicity of the drug and the damage to the cortical and sub-cortical
systems, such as the serotonergic system, stemming from the psychoses
initiated by using MA (Sekine et al., 2006).
Elevated complexity in EEG patterns mainly in the left centro-
parietal areas was found in those MA abusers that participated in sexual
intercourse during their MA binges. Previous reports have suggested
that the parietal regions are associated with erotic, especially visual,
stimuli (Montorsi et al., 2003; Mouras et al., 2003). Our results suggest
that the sexual intercourse that occurs concurrently with MA abuse is, to
some extent, associated with relatively intact brain function, rather than
a simple recreational behavior. Further investigation is required to
examine not only former abusers but also current and potential abusers
to confirm the measure's diagnostic power.
Fig. 4. ApEn values of MA abusers correlated with the presence of nalbuphine abuse. Our findings must be interpreted in light of the limitations of this
The asterisk (*) indicates significant differences in ApEn values (p b 0.05). study. First, the precise mental processes during resting remain
Please cite this article as: Yun, K., et al., Decreased cortical complexity in methamphetamine abusers, Psychiatry Research: Neuroimaging
(2012), doi:10.1016/j.pscychresns.2011.07.009
6. 6 K. Yun et al. / Psychiatry Research: Neuroimaging xxx (2012) xxx–xxx
essentially uncontrolled. Different mental states might have influ- Hornero, R., Espino, P., Alonso, A., Lopez, M., 1999. Estimating complexity from EEG
background activity of epileptic patients. IEEE Engineering in Medicine and Biology
enced the measured entropy. This is probably the main limitation for Magazine 18, 73–79.
utilizing resting state dynamics. However, random mind wandering Jeong, J., Kim, S.Y., Han, S.H., 1998. Non-linear dynamical analysis of the EEG in
could be canceled out in between-subjects level statistics. Second, the Alzheimer's disease with optimal embedding dimension. Electroencephalography
and Clinical Neurophysiology 106, 220–228.
duration and amounts of MA use were determined by retrospective Johanson, C.E., Frey, K.A., Lundahl, L.H., Keenan, P., Lockhart, N., Roll, J., Galloway, G.P.,
self-reports, which have limited validity and precision. Third, the Koeppe, R.A., Kilbourn, M.R., Robbins, T., 2006. Cognitive function and nigrostriatal
decision to study only males, who likely are more susceptible than are markers in abstinent methamphetamine abusers. Psychopharmacology 185, 327–338.
Kato, Y., Muramatsu, T., Kato, M., Shibukawa, Y., Shintani, M., Yoshino, F., 2006. Cortical
women to the neurotoxic effects of methamphetamine, limited the reorganization and somatic delusional psychosis: a magnetoencephalographic
generalizability of our findings to men only. Fourth, detailed study. Psychiatry Research: Neuroimaging 146, 91–95.
neuropsychological testing was not conducted, and therefore the Keshavan, M.S., Cashmere, J.D., Miewald, J., Yeragani, V.K., 2004. Decreased nonlinear
complexity and chaos during sleep in first episode schizophrenia: a preliminary
functional correlates of the altered EEG complexity patterns detected
report. Schizophrenia Research 71, 263–272.
in this study could not be assessed. Finally, although we excluded Kim, S.J., Lyoo, I.K., Hwang, J., Chung, A., Hoon Sung, Y., Kim, J., Kwon, D.H., Chang, K.H.,
subjects with extensive nicotine or alcohol use, the methamphet- Renshaw, P.F., 2005a. Prefrontal grey-matter changes in short-term and long-term
amine abusers likely consumed larger amounts of other drugs during abstinent methamphetamine abusers. The International Journal of Neuropsycho-
pharmacology 9, 221–228.
their lifetimes than did the comparison group, thereby confounding Kim, S.J., Lyoo, I.K., Hwang, J., Sung, Y.H., Lee, H.Y., Lee, D.S., Jeong, D.U., Renshaw, P.F.,
our ability to attribute the causes of hypoactivity to MA alone. 2005b. Frontal glucose hypometabolism in abstinent methamphetamine users.
Nevertheless, the complexity analyses from our EEG screenings might Neuropsychopharmacology 30, 1383–1391.
Kita, T., Wagner, G.C., Nakashima, T., 2003. Current research on methamphetamine-
provide insights into how information is abnormally processed in the induced neurotoxicity: animal models of monoamine disruption. Journal of
cortical networks of MA abusers. This study suggests the possibility of Pharmacological Sciences 92, 178–195.
dynamical complexity measures like ApEn as a potential, supplemen- Kraemer, T., Maurer, H.H., 2002. Toxicokinetics of amphetamines: metabolism and
toxicokinetic data of designer drugs, amphetamine, methamphetamine, and their
tary quantitative diagnostic tool for MA abuse. N-alkyl derivatives. Theraputic Drug Monitoring 24, 277–289.
Kulsudjarit, K., 2004. Drug problem in Southeast and Southwest Asia. Annals of the New
York Academy of Sciences 1025, 446–457.
Acknowledgments Lautenschlager, N.T., Forstl, H., 2001. Organic psychosis: insight into the biology of
psychosis. Current Psychiatry Reports 3, 319–325.
The authors thank Mr. Seung Min Shin for his technical assistance of Lin, M.A., Chan, H.L., Fang, S.C., 2005. Linear and nonlinear EEG indexes in relation to the
severity of coma. Engineering in Medicine and Biology Society, 2005. IEEE-EMBS
EEG recordings. This work was supported by the Korea Science and 2005. 27th Annual International Conference, pp. 4580–4583.
Engineering Foundation (KOSEF) grant funded by the Korea government London, E.D., Simon, S.L., Berman, S.M., Mandelkern, M.A., Lichtman, A.M., Bramen, J.,
(MOST) (No. R01-2007-000-21094-0 and No. M10644000013- Shinn, A.K., Miotto, K., Learn, J., Dong, Y., 2004. Mood disturbances and regional
cerebral metabolic abnormalities in recently abstinent methamphetamine abusers.
06N4400-01310). Archives of General Psychiatry 61, 73–84.
McCann, U.D., Wong, D.F., Yokoi, F., Villemagne, V., Dannals, R.F., Ricaurte, G.A., 1998.
Reduced striatal dopamine transporter density in abstinent methamphetamine and
Appendix A. Supplementary data methcathinone users: evidence from positron emission tomography studies with
[11C] WIN-35,428. The Journal of Neuroscience 18, 8417–8422.
McCann, U.D., Kuwabara, H., Kumar, A., Palermo, M., Abbey, R., Brasic, J., Ye, W.,
Supplementary data to this article can be found online at doi:10.
Alexander, M., Dannals, R.F., Wong, D.F., 2008. Persistent cognitive and dopamine
1016/j.pscychresns.2011.07.009. transporter deficits in abstinent methamphetamine users. Synapse 62, 91–100.
Micheloyannis, S., Pachou, E., Stam, C.J., Breakspear, M., Bitsios, P., Vourkas, M., Erimaki,
S., Zervakis, M., 2006. Small-world networks and disturbed functional connectivity
References in schizophrenia. Schizophrenia Research 87, 60–66.
Montorsi, F., Perani, D., Anchisi, D., Salonia, A., Scifo, P., Rigiroli, P., Deho, F., De Vito, M.L.,
Abasolo, D., Hornero, R., Espino, P., Poza, J., Sanchez, C.I., de la Rosa, R., 2005. Analysis of Heaton, J., Rigatti, P., 2003. Brain activation patterns during video sexual
regularity in the EEG background activity of Alzheimer's disease patients with stimulation following the administration of apomorphine: results of a placebo-
Approximate Entropy. Clinical Neurophysiology 116, 1826–1834. controlled study. European Urology 43, 405–411.
Bob, P., Susta, M., Prochazkova-Vecerova, A., Kukleta, M., Pavlat, J., Jagla, F., Raboch, J., Mouras, H., Stoleru, S., Bittoun, J., Glutron, D., Pelegrini-Issac, M., Paradis, A.L., Burnod,
2006. Limbic irritability and chaotic neural response during conflicting Stroop task Y., 2003. Brain processing of visual sexual stimuli in healthy men: a functional
in the patients with unipolar depression. Physiological Research 55, S107–S112. magnetic resonance imaging study. NeuroImage 20, 855–869.
Breakspear, M., Terry, J.R., Friston, K.J., Harris, A.W., Williams, L.M., Brown, K., Brennan, Newlin, D.B., Wong, C.J., Stapleton, J.M., London, E.D., 2000. Intravenous cocaine decreases
J., Gordon, E., 2003. A disturbance of nonlinear interdependence in scalp EEG of cardiac vagal tone, vagal index (derived in Lorenz space), and heart period complexity
subjects with first episode schizophrenia. NeuroImage 20. (approximate entropy) in cocaine abusers. Neuropsychopharmacology 23, 560–568.
Burioka, N., Cornelissen, G., Maegaki, Y., Halberg, F., Kaplan, D.T., Miyata, M., Fukuoka, Newton, T.F., Cook, I.A., Kalechstein, A.D., Duran, S., Monroy, F., Ling, W., Leuchter, A.F.,
Y., Endo, M., Suyama, H., Tomita, Y., 2005a. Approximate entropy of the 2003. Quantitative EEG abnormalities in recently abstinent methamphetamine
electroencephalogram in healthy awake subjects and absence epilepsy patients. dependent individuals. Clinical Neurophysiology 114, 410–415.
Clinical EEG and Neuroscience 36, 188–193. Newton, T.F., Kalechstein, A.D., Hardy, D.J., Cook, I.A., Nestor, L., Ling, W., Leuchter, A.F.,
Burioka, N., Miyata, M., Corn?issen, G., Halberg, F., Takeshima, T., Kaplan, D.T., Suyama, H., 2004. Association between quantitative EEG and neurocognition in methamphet-
Endo, M., Maegaki, Y., Nomura, T., 2005b. Approximate entropy in the electroenceph- amine-dependent volunteers. Clinical Neurophysiology 115, 194–198.
alogram during wake and sleep. Clinical EEG and Neuroscience 36, 21. Office of Applied Studies, 2005. Overview of Findings from the 2004 National Survey on
Chen, C.K., Lin, S.K., Sham, P.C., Ball, D., Loh, E.W., Hsiao, C.C., Chiang, Y.L., Ree, S.C., Lee, C.H., Drug Use and Health. (Publication No. 05-4061) Substance Abuse and Mental
Murray, R.M., 2003. Pre-morbid characteristics and co-morbidity of methamphet- Health Services Administration, DHHS, Rockville. MD.
amine users with and without psychosis. Psychological Medicine 33, 1407–1414. Paulus, M., 2002. Behavioral and functional neuroimaging evidence for prefrontal
Chung, H., Park, M., Hahn, E., Choi, H., Choi, H., Lim, M., 2004. Recent trends of drug dysfunction in methamphetamine-dependent subjects. Neuropsychopharmacology
abuse and drug-associated deaths in Korea. Annals of the New York Academy of 26, 53–63.
Sciences 1025, 458–464. Paulus, M.P., Braff, D.L., 2003. Chaos and schizophrenia: does the method fit the
Ciraulo, D.A., Piechniczek-Buczek, J., Iscan, E.N., 2003. Outcome predictors in substance madness? Biological Psychiatry 53, 3–11.
use disorders. The Psychiatric Clinics of North America 26, 381–409. Pincus, S.M., 1991. Approximate entropy as a measure of system complexity.
D'Astous, M., Mickley, K.R., Dluzen, D.E., Di Paolo, T., 2005. Differential protective Proceedings of the National Academy of Sciences 88, 2297–2301.
properties of estradiol and tamoxifen against methamphetamine-induced nigros- Pincus, S.M., 1994. Greater signal regularity may indicate increased system isolation.
triatal dopaminergic toxicity in mice. Neuroendocrinology 82, 111–120. Mathematical Biosciences 122, 161–181.
Dluzen, D.E., Anderson, L.I., Pilati, C.F., 2002. Methamphetamine-gonadal steroid Pincus, S., 1995. Approximate entropy (ApEn) as a complexity measure. Chaos 5, 110–117.
hormonal interactions: effects upon acute toxicity and striatal dopamine Pincus, S.M., Goldberger, A.L., 1994. Physiological time-series analysis: what does
concentrations. Neurotoxicology and Teratology 24, 267–273. regularity quantify? American Journal of Physiology. Heart and Circulatory
Farrell, M., Marsden, J., Ali, R., Ling, W., 2002. Methamphetamine: drug use and Physiology 266, 1643–1656.
psychoses becomes a major public health issue in the Asia Pacific region. Addiction Radhakrishnan, N., Gangadhar, B.N., 1998. Estimating regularity in epileptic seizure
97, 771–772. time-series data. Engineering in Medicine and Biology Magazine, IEEE 17, 89–94.
Fingelkurts, A.A., Rytsala, H., Suominen, K., Isometsa, E., Kahkonen, S., 2007. Impaired Ricaurte, G.A., Schuster, C.R., Seiden, L.S., 1980. Long-term effects of repeated
functional connectivity at EEG alpha and theta frequency bands in major methylamphetamine administration on dopamine and serotonin neurons in the
depression. Human Brain Mapping 28 (3), 247–261. rat brain: a regional study. Brain Research 193, 153–163.
Please cite this article as: Yun, K., et al., Decreased cortical complexity in methamphetamine abusers, Psychiatry Research: Neuroimaging
(2012), doi:10.1016/j.pscychresns.2011.07.009
7. K. Yun et al. / Psychiatry Research: Neuroimaging xxx (2012) xxx–xxx 7
Roschke, J., Mann, K., Fell, J., 1994. Nonlinear EEG dynamics during sleep in depression Tononi, G., Edelman, G.M., 1998. Consciousness and complexity. Science 282, 1846–1851.
and schizophrenia. The International Journal of Neuroscience 75, 271–284. Tononi, G., Sporns, O., Edelman, G.M., 1994. A measure for brain complexity: relating
Sattler, J.M., 2001. Assessment of Children: Cognitive Applications, Fourth Edition. PRO-ED functional segregation and integration in the nervous system. Proceedings of the
Inc. National Academy of Sciences 91, 5033–5037.
Seivewright, N., 2000. Disorders relating to the use of amphetamine and cocaine. New Tononi, G., Edelman, G.M., Sporns, O., 1998. Complexity and coherency: integrating
Oxford Textbook of Psychiatry. Oxford University, New York, pp. 531–534. information in the brain. Trends in Cognitive Sciences 2, 474–484.
Sekine, Y., Iyo, M., Ouchi, Y., Matsunaga, T., Tsukada, H., Okada, H., Yoshikawa, E., Villemagne, V., Yuan, J., Wong, D.F., Dannals, R.F., Hatzidimitriou, G., Mathews, W.B.,
Futatsubashi, M., Takei, N., Mori, N., 2001. Methamphetamine-related psychiatric Ravert, H.T., Musachio, J., McCann, U.D., Ricaurte, G.A., 1998. Brain dopamine
symptoms and reduced brain dopamine transporters studied with PET. The neurotoxicity in baboons treated with doses of methamphetamine comparable to
American Journal of Psychiatry 158, 1206–1214. those recreationally abused by humans: evidence from [11C] WIN-35,428 positron
Sekine, Y., Minabe, Y., Ouchi, Y., Takei, N., Iyo, M., Nakamura, K., Suzuki, K., Tsukada, H., emission tomography studies and direct in vitro determinations. The Journal of
Okada, H., Yoshikawa, E., 2003. Association of dopamine transporter loss in the Neuroscience 18, 419–427.
orbitofrontal and dorsolateral prefrontal cortices with methamphetamine-related Volkow, N.D., Chang, L., Wang, G.J., Fowler, J.S., Franceschi, D., Sedler, M., Gatley, S.J.,
psychiatric symptoms. The American Journal of Psychiatry 160, 1699–1701. Miller, E., Hitzemann, R., Ding, Y.S., 2001a. Loss of dopamine transporters in
Sekine, Y., Ouchi, Y., Takei, N., Yoshikawa, E., Nakamura, K., Futatsubashi, M., Okada, H., methamphetamine abusers recovers with protracted abstinence. The Journal of
Minabe, Y., Suzuki, K., Iwata, Y., Tsuchiya, K.J., Tsukada, H., Iyo, M., Mori, N., 2006. Neuroscience 21, 9414.
Brain serotonin transporter density and aggression in abstinent methamphetamine Volkow, N.D., Chang, L., Wang, G.J., Fowler, J.S., Leonido-Yee, M., Franceschi, D., Sedler,
abusers. Archives of General Psychiatry 63, 90–100. M.J., Gatley, S.J., Hitzemann, R., Ding, Y.S., 2001b. Association of dopamine
Spitzer, R.L., Williams, J.B.W., Gibbon, M., 1994. Structured Clinical Interview for DSM-IV transporter reduction with psychomotor impairment in methamphetamine
(SCID). Biometrics Research Department, New York State Psychiatric Institute, New York. abusers. The American Journal of Psychiatry 158, 377–382.
Sporns, O., Tononi, G., Edelman, G.M., 2000. Connectivity and complexity: the relationship Wada, K., Fukui, S., 1990. Relationship between years of methamphetamine use and
between neuroanatomy and brain dynamics. Neural Networks 13, 909–922. symptoms of methamphetamine psychosis. Arukoru Kenkyuto Yakubutsu Ison 25,
Taillard, J., Philip, P., Coste, O., Sagaspe, P., Bioulac, B., 2003. The circadian and 143–158.
homeostatic modulation of sleep pressure during wakefulness differs between Wagner, G.C., Ricaurte, G.A., Johanson, C.E., Schuster, C.R., Seiden, L.S., 1980.
morning and evening chronotypes. Journal of Sleep Research 12, 275–282. Amphetamine induces depletion of dopamine and loss of dopamine uptake sites
Thomasson, N., Pezard, L., Boyer, P., Renault, B., Martinerie, J., 2002. Nonlinear EEG in caudate. Neurology 30, 547–550.
changes in a 48-hour cyclic manic-depressive patient. Nonlinear Dynamics, Yeom, T.H., Park, Y.S., Oh, K.J., Kim, J.K., Lee, Y.H., 1992. K-WAIS Application Principles.
Psychology, and Life Sciences 6, 259–267. Korea Guidance, Seoul.
Thompson, P.M., Hayashi, K.M., Simon, S.L., Geaga, J.A., Hong, M.S., Sui, Y., Lee, J.Y., Toga,
A.W., Ling, W., London, E.D., 2004. Structural abnormalities in the brains of human
subjects who use methamphetamine. The Journal of Neuroscience 24, 6028–6036.
Please cite this article as: Yun, K., et al., Decreased cortical complexity in methamphetamine abusers, Psychiatry Research: Neuroimaging
(2012), doi:10.1016/j.pscychresns.2011.07.009
8. Supporting Online Material
Decreased Cortical Complexity in Methamphetamine abusers
Kyongsik Yun*, Hee-Kwon Park*, Do-Hoon Kwon, Yang-Tae Kim
Sung-Nam Cho, Hyun-Jin Cho, Bradley S. Peterson, Jaeseung Jeong
Materials and methods
Approximate Entropy
EEG can be reconstructed as phase space representation using the method of delays.
In this representation, the state at each point in time is represented by a vector generated by
taking successive amplitudes separated by a time lag tau. This reconstruction of the
underlying dynamics is the first step of all techniques of phase space analysis. The
geometrical properties of the trajectories in the phase space can then be expressed
quantitatively using nonlinear measures. Approximate entropy can be represented by
difference between geometrical property of m dimensional phase space trajectory and that of
m+1 dimensional phase space trajectory.
In detail, ApEn is derived from the correlation integral Cim (r ) , which represents the
number of points within a distance r from the ith point when the signal is embedded in an m-
dimensional space:
N - ( m -1)
m -1
C (r ) = ( N - (m - 1))
i å Q(r - Xi - Xj )
j =1
(1)
Where Q(t ) is the Heaviside function (if t ³ 0, Q(t ) = 1 ; if t<0, Q(t ) =0) and Xi and Xj are
vectors constructed from the time series [ x(1), x(2), . . . , x(N)] as
1
9. Xi = {x(i ), x(i + t ),..., x(i + (m - 1)t )}
Xj = {x( j ), x( j + t ),..., x( j + (m - 1)t )} (2)
(i, j = 1,2,...N - (m - 1)t )
where t is the time lag at which the mutual information between consecutive samples
becomes negligible. The ApEn statistic is defined as follows:
ApEn(m, r ) = Qm (r ) - Qm +1 (r ) (3)
N - ( m -1)
Q m (r ) = ( N - (m - 1)) -1 å ln C i
m
(r ) (4)
i =1
Entropy can be viewed as a measure of disorder, larger values corresponding more disorder,
randomness, or complexity (Williams, 1997). We set the distance, r, to be 0.2 times the SD of
the original data series to produce reasonable statistical validity of ApEn (Pincus, 1991). We
tested embedding dimension from 2 to 6. We used an embedding dimension of 2 (ie, m = 2)
(Peitgen et al., 1992) in which ApEn differences between groups are maximum. The
parameter choices of m = 2 and r = 0.2 SD in the ApEn specification are standard choices and
widely applied in diverse settings. To embed the time series in state space, we used the
concept of time lag. These values were chosen as the time lags that were determined for each
time series as the lag at which the first minimum of the mutual information in the EEG. ApEn
was computed with a software package (MATLAB 7.0; The MathWorks, Inc).
Surrogate data
The surrogate data are a randomized sequence of the original data having the same linear
properties and the significant differences between the ApEn of original EEG data and their
surrogate data indicate that the original EEG sequences have a nonlinear structure within the
patterns. Nonlinear indexes such as the ApEn are computed for several surrogate data series.
Their values are compared with that assumed by the nonlinear index computed for the
2
10. original data (Theiler et al., 1992).
No statistical difference indicates that the original time series were generated from a
linear process, since random shuffling of the original time series, which is surrogate data,
does not change the linear properties of the original data. On the contrary, the statistically
significant difference in ApEn between the original and surrogate data explains that the
original signal contains the nonlinear properties.
To test for a statistical significance of difference (ie, the s of Theiler et al. (Theiler et al.,
1992)) in ApEn between the original and the surrogate data, 10 surrogate data series were
generated to match each original signal. Let ApEnorig be the ApEn of the original data, and let
ApEnsurr be the ApEn of the 10 surrogate series (i = 1,…,10). The mean and SD of ApEnsurr (i
= 1,…,10) are estimated as ApEnsurr and SDApEn_surr (Theiler et al., 1992) then is computed as
follows:
ApEnorig - < ApEnsurr >
Z= (5)
SDApEn _ surr
This statistic represents the number of SDs ( s ) distant from ApEnorig. It follows a Student t
test distribution with 9 degrees of freedom ( t9[1- a /2]). For a =0.05, the critical value of t
is 2.26. Accordingly, when the s of Theiler et al. (Theiler et al., 1992) is > 2.26, the null
hypothesis is rejected at the 5% probability level, and the original data are considered to
contain nonlinear features.
3
11. Results
Power spectrum analysis
The reformatted data were than processed using a fast Fourier transform (MATLAB 7.0; The
MathWorks, Inc) to obtain relative power values in 4 standard frequency bands, delta (0.5–4
Hz), theta (4–8 Hz), alpha (8–12 Hz), and beta (12–30 Hz).
Relative power values were obtained across the 4 standard EEG bands. Areas in
prefrontal, frontal, and temporal regions were decreased (Fp1, F8, T5; p<0.05) in MA-
dependent group in delta band. Other EEG bands indicate no significantly different power
values between MA-dependent and control groups.
Surrogate data analysis
The test of Theiler et al. (Theiler et al., 1992) was performed for each data series
separately to test for nonlinearity (FigS1). The mean (± standard error) values of the test of
Theiler et al. (Theiler et al., 1992) were 41.3 ± 5.3 for ApEn of MA-dependent subjects, 44.2
± 6.3 for ApEn of high-dose MA-dependent subjects, 35.5 ± 6.6 for ApEn of low-dose MA-
dependent subjects. The mean values of the control group were 2.33 ± 0.5. The signals of all
channels of MA group show clear evidence for nonlinearity) Whereas signals of 8 channels
of control group (Fp2, C3, C4, F8, T3, T4, T5, T6) failed the test signals of 8 other channels
of control group (Fp1, F3, F4, P3, P4, F7, O1, O2) showed nonlinearity. It shows relatively
significant nonlinear feature in MA-dependent subjects. The signals of all channels of MA
group show clear evidence for nonlinearity although signals for only 8 channels show
significant nonlinear feature in the control group.
Correlation with abstinence period
4
12. We could not find the correlation between abstinence period and EEG analyses such
as ApEn and power spectrum. Although Kim et al. (Kim et al., 2005) reported that long-term
abstinence could improve prefrontal grey-matter, average duration of long-term abstinence
period was over 30 months in that study. In case of our study, the longest abstinence period
was three months, and the relatively short period of abstinence maybe could not show enough
recovery of electrophysiologic function.
5
13. Discussion
Brecht and his colleagues (Brecht et al., 2000) examined the possible predictors of
that could indicate how long MA abusers could wait before relapsing back into abusing the
drug. They found that an abuser experienced a shorter time period before relapse if the
individual had started using MA at an older age, and age was found to be a significant
predictor in that experiment. In our study, the ApEn values of the left frontal areas exhibit
negative correlations with the ages of subjects’ first intakes of the drug. This preliminary
finding indicates that MA abusers who were older when first exposed to MA have elevated
levels of susceptibility to reductions in their EEG complexities. It is possible that the
dopamine systems is more severely damaged by MA as individual ages since these
neurotransmitter systems are reported to weaken during the course of the normal aging
process. This might compound the effects of the drug as well (Sheline et al., 2002). The
association of ApEn values with the severity of MA abuse and the age at which a user first
uses the drug suggests that EEG complexity measures should be further investigated.
The ApEn revealed decreases in the global cortical areas in those MA abusers that
had drug-related criminal records. Drug-related criminality is frequently associated with MA
abuse (Keene, 2005; Sindelar & Olmstead, 2006). To investigate certain social aspects, such
as the levels of sociality and aggression in MA abusers, we examined the other criminal
records unrelated to substance abuse of our subjects in the patient group. We found that the
general criminality of MA abusers was associated with decreased complex dynamics in the
left fronto-central and right centro-parietal areas that are known to be associated with
increased hostility (Fallon et al., 2004) and more frequent and intense sexual behaviors
(Mouras, 2006). The serotonin transporter likely changes during an individual’s dependence
on MA, leading to elevated levels of aggression even in currently former abusers (Sekine et
6
14. al., 2006). However, the ApEn analysis revealed increased complexity in these areas
correlated with sexual activity. Thus, this discrepancy suggests that further investigation is
necessary to identify the specific processes of those centro-parietal areas involved in sexual
activity and criminality.
In regards to our observations that showed heavy drinking and smoking habits in
MA abusers, they were associated with EEG complexity in the left frontal and temporal
regions which was consistent with a previous report that stated smoking increased the
psycho-toxic effect of and sensitivity to MA, particularly in the inhibitions of locomotor
sensitization (Kuribara, 1999). We should also note that co-morbid use of nalbuphine
minimized the ApEn decrease that was induced by MA abuse, or this co-morbid use increased
the ApEn in most cortical areas of abusers compared to the healthy control subjects. This is
consistent with previous studies performed on animals that reported that stimulation of the
opioid receptors plays an inhibitory role in MA-induced and self-injurious behaviors (Mori et
al., 2006). The results implicate that ApEn may be able to detect the outcomes of cortical
dynamics produced by an interplay between nalbuphine and methamphetamine. Alternatively,
nalbuphine may offer some protection against MA-induced reductions in cortical complexity
and can play a potential and therapeutic role in treating MA-inducing pathological changes.
7
15. Supporting References
Brecht, M. L., von Mayrhauser, C., & Anglin, M. D. 2000. Predictors of relapse after
treatment for methamphetamine use. J Psychoactive Drugs 32, 211-20.
Fallon, J. H., Keator, D. B., Mbogori, J., Turner, J., & Potkin, S. G. 2004. Hostility
differentiates the brain metabolic effects of nicotine. Brain Res Cogn Brain Res 18,
142-8.
Keene, J. 2005. A case-linkage study of the relationship between drug misuse, crime, and
psychosocial problems in a total criminal justice population. Addiction Research &
Theory 13, 489-502.
Kim, S. J., Lyoo, I. K., Hwang, J., Chung, A., Hoon Sung, Y., Kim, J., Kwon, D. H., Chang,
K. H., & Renshaw, P. F. 2005. Prefrontal grey-matter changes in short-term and long-
term abstinent methamphetamine abusers. The International Journal of
Neuropsychopharmacology 9, 221-228.
Kuribara, H. 1999. Does nicotine modify the psychotoxic effect of methamphetamine?
Assessment in terms of locomotor sensitization in mice. J Toxicol Sci 24, 55-62.
Mori, T., Ito, S., Kita, T., Narita, M., Suzuki, T., & Sawaguchi, T. 2006. Effects of mu-, delta-
and kappa-opioid receptor agonists on methamphetamine-induced self-injurious
behavior in mice. Eur J Pharmacol.
Mouras, H. 2006. Neuroimaging Techniques as a New Tool to Study the Neural Correlates
Involved in Human Male Sexual Arousal. Current Medical Imaging Reviews 2, 71-77.
Peitgen, H.-O., Jurgens, H., & Saupe, D. (1992). Chaos and fractals : new frontiers of science.
New York: Springer-Verlag.
Pincus, S. M. 1991. Approximate entropy as a measure of system complexity. Proceedings of
National Academy of Sciences 88, 2297-301.
8
16. Sekine, Y., Ouchi, Y., Takei, N., Yoshikawa, E., Nakamura, K., Futatsubashi, M., Okada, H.,
Minabe, Y., Suzuki, K., Iwata, Y., Tsuchiya, K. J., Tsukada, H., Iyo, M., & Mori, N.
2006. Brain serotonin transporter density and aggression in abstinent
methamphetamine abusers. Archives of General Psychiatry 63, 90-100.
Sheline, Y. I., Mintun, M. A., Moerlein, S. M., & Snyder, A. Z. 2002. Greater loss of 5-
HT(2A) receptors in midlife than in late life. Am J Psychiatry 159, 430-5.
Sindelar, J. L., & Olmstead, T. A. 2006. 8 Illicit drugs and drug-related crime. The Elgar
Companion to Health Economics.
Theiler, J., Eubank, S., Longtin, A., Galdrikian, B., & Doyne Farmer, J. 1992. Testing for
nonlinearity in time series: the method of surrogate data. Physica D: Nonlinear
Phenomena 58, 77-94.
Williams, G. P. (1997). Chaos theory tamed. Washington, D.C.: Joseph Henry Press.
9
17. Supporting Figures
FigS1. Surrogate data analysis. A global test of nonlinearity was carried out by a Wilcoxon
matched-pairs signed rank test comparing ApEn values computed on the original data paired
with the corresponding average ApEn values from the matching surrogate data series. Larger
sigma values indicate higher nonlinearity.
10
18. FigS2. Correlation with criminal records. The ApEn values of the EEG correlated with
drug-related criminal records. Asterisk (*) indicates significant difference in the ApEn values.
11
19. FigS3. Correlation with smoking. The ApEn values of MA abusers correlated with
heavy/light smoking. Heavy smoking is defined as more than or equal to 2 packs/day.
Asterisk (*) indicates significant difference in the ApEn values.
12