Open Access
Add to BibSonomy
Abstract
Memory shortness, verbal influence, and disturbed attention are a few of the cognitive dysfunctions reported by individuals of bipolar disorder in depression phase (BD-D). As neuroimaging modalities can investigate such responses, therefore neuroimaging methods can be used to assist the diagnosis of bipolar disorder (BD). Functional near-infrared spectroscopy (fNIRS) is a neural imaging method that is proved to be prominent in the diagnosis of psychiatric disorders. It is the desired method because of its feasible setup, high resolution in time, and its partial resistance to head movements. This study aims to investigate the brain activity in subjects of BD-D during cognitive tasks compared to the healthy controls. A decreased activation level is expected in individuals of BD-D as compared to the healthy controls. This study aims to find new methods and experimental paradigms to assist in the diagnosis of bipolar depression. Participants of BD-D and healthy controls (HC) performed four cognitive tasks including verbal fluency task (VFT), symbol working memory task (symbol check), attention task (spotter) and multiple cognitive task (code break). fNIRS was used to measure levels of oxy-hemoglobin (HbO) representing the brain activity. The generalized linear model (GLM) method was used to estimate the hemodynamic response related to the task. The wavelet transform coherence (WTC) method was used to calculate the intra-hemispheric functional connectivity. We also analyzed the correlation between hemodynamic response and scores of psychiatric disorders. Results showed decreased levels of HbO in BD-D groups compared to the HC, indicating lower activity, during the tasks except for spotter. The difference between BD-D and HC was significant during VFT, symbol check and code break. Group difference during symbol working memory was significant both in brain activity and connectivity. Meanwhile, the individual brain activity during working memory is more related to the illness degree. Lower activity in BD-D reflects unspecific dysfunctions. Compared with other cognitive tasks, the single-trial symbol-check task may be more suitable to help the diagnosis of bipolar depression.
© 2022 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
Corrections
3 March 2022: A typographical correction was made to the author affiliations.
1. Introduction
Bipolar disorder (BD) is characterized by bipolar episodes of mania or hypomania and depression, and manifests as a chronic mental disorder with recurrent episodes. BD often starts in adolescence or early adulthood but remain undiagnosed for an extended period of time [1]. BD has been ranked as one of the 20 leading medical causes of disability (WHO,2011), and is the 9th leading cause of Disability-Adjusted Life-Years in people between 15 and 45 years of age. Most of the diagnosis cannot be made on the first visit. Therefore, it is very important to have effective auxiliary diagnostic tools. In clinical practice, it has been shown that a delay of diagnosis has a negative influence on the progression of the disorder, whereas an early and appropriate treatment has beneficial effects [2]. Therefore, further exploration of the pathogenesis and biomarkers of BD is of great significance for the clinical diagnosis and treatment. Prominent cognitive deficits, include planning, operational memory, attention, problem-solving, inhibition control, and psychological flexibility, especially altered reaction time, verbal and visual memory and executive function, have been documented in bipolar disorder [3–5]. Biological markers linking neurophysiological and behavioral symptoms of psychosis have been explored. In a comprehensive review [6], describe recent findings of fNIRS studies in schizophrenia research, which showed decreased activation of prefrontal and frontoparietal cortices during various neuropsychological tasks such as: verbal fluency task, tower of London, continuous performance test and kana Stroop task.
In terms of bipolar cognitive deficits, many cranial functional magnetic resonance (fMRI) studies have provided some insights. The dorsolateral prefrontal cortex (dlPFC) region has been suggested to play a critical superordinate role in regulating cognitive control being involved in various cognitive tasks where executive functions are recruited [7]. Impaired patients of BD exhibited reduced activation compared to cognitively normal patients during a verbal N-back working memory (WM) task. The task-related brain regions were inferior parietal gyri (IPG), dlPFC, and dorsomedial prefrontal cortex (dmPFC). Impaired patients exhibited less deactivation of the default mode network (DMN) than cognitively normal patients across orbitofrontal cortex (OFC), posterior cingulate gyrus (PCG), and fusiform gyrus (FFG) [8].
Functional near infrared spectroscopy (fNIRS) is a promising neuroimaging technology for brain function and blood oxygenation detection [9,10]. fNIRS places fewer limitations on the subjects and environment as compared to other functional neuroimaging methods such as fMRI and Positron Emission Tomography (PET) [11,12]. fNIRS researches range from the phenomenological characterization of psychiatric disorders, descriptions of lifetime developmental aspects, clinical monitoring, treatment effects, and genetic influences on neuroimaging data [13]. The development of multichannel NIRS systems has allowed the measurement of interactions between brain regions to derive NIRS based connectivity [14–17]. Therefore, fNIRS is an appropriate tool to study the brain functions of patients with BD and may be used in clinical diagnosis.
Although there are many studies on brain functions of psychiatric disorders detected by fNIRS [18,19], only a few studies focused on bipolar disorder in depression phase (BD-D). A multicenter, collaborative NIRS study reported that the accuracy of left frontal hemodynamic responses in the diagnosis of bipolar disorder is 77% [20]. In the clinical diagnosis of BD, verbal fluency task (VFT) is most frequently used [21,22]. The participants are asked to use the given word to group words as many as they can or list as many items as possible that belong to the same category. However, VFT only covers a restricted aspect of executive functioning, which would be inadequate in delineating and differentiating complex psychiatric disorders including bipolar depression [23]. Meanwhile, the process control and behavioral recordings of VFT are difficult, according to the response of clinical outpatient doctors. For past few years the application of fNIRS in clinical diagnosis has an increasing trend, therefore, we need to explore more reliable and effective tasks to cooperate with brain function detection for the diagnosis of BD.
The THINC-integrated tool (THINC-it) was released at the 2016 annual meeting of the European Society of Neuropsychopharmacology (ECNP) (website: https:// thinc.progress.im/en). The THINC-it tool is free of charge, digitalized, patient–administered, and has been validated as a screening tool in adults with Major Depressive Disorder. The THINC-it tool can be used at the point of care, as a repeated measure across time, and provides the end-user with an easy-to-translate assessment in adults with Major Depressive Disorder. The THINC-it includes Spotter, Symbol Check, Codebreaker, Trails, and PDQ-5-D. Spotter is used to investigate attention and executive function. The Symbol Check evaluates working memory and executive function. The Codebreaker focuses on executive functions, processing speed, and attention/concentration. THINC-it is recommended for early detection of cognitive impairment in mood disorders including BD [30]. Studies report high sensitivity of THINC-IT in the evaluation of cognitive function in BD patients [31]. Therefore, we have chosen THINC-it as the experimental tool in this study.
Until now, not many studies have investigated the brain dysfunction in subject of BD-D during cognitive tasks except VFT. This study employs fNIRS to measure the brain activity of subjects of BD-D and compare them with those of healthy control group (HC) during multi-cognitive tasks. The study aims to explore the influence of BD on cognitive functions from the perspective of cortex hemodynamic alternations. The hemodynamic activation of the prefrontal cortex (PFC) during cognitive tasks will be observed in the both groups, and significantly lower frontotemporal activity measured by HbO levels in BD-D patients.
2. Material and methods
2.1 Subjects
Thirty-three participants aged between 14 and 35 years, diagnosed with BD-D by senior doctors, (according to DSM-5 criteria and confirmed on MINI questionnaire), were recruited from the hospital, and labeled as BD-D group. 22 healthy participants (with no history of any mental disorder) aged between 16 and 25 years were recruited and labeled as HC group. For the BD-D group, participants had to score >16 points on the Hamilton Depression Rating Scale (HAMD). The two groups were comparable in age, gender, and years of education (all p > 0.1). Results of 5 subjects (4 in BD and 1 in HC) were excluded from the following data analysis due to insufficient signal quality caused by bad scalp contacts of the optical probes and severe muscle artifacts. The final HC groups were 21 subjects (aged 22.7 ± 2.9 years) including 11 women. The final BD-D groups were 29 subjects (aged 23.1 ± 4.7 years) including 17 women. All the subjects were right-handed. This study was approved by Clinical Research Ethics Committee of the First Affiliated Hospital, College of Medicine, Zhejiang University (IIT20210036C-R1).
2.2 Protocol
The experiment includes 4 different cognitive tasks. First is the classical VFT task which is normally used in clinical diagnosis. At the beginning of the task, the screen has a 30 s blank period, and participants need to repeatedly count from one to five until the task starts. Then, in VFT task, the participants are asked to speak out names of as many animals as possible, and the correct number of enumerating animal names was counted. After 60 s, the screen returns to blank, and the participants repeat the counting as done in the pre-task period. The other three tasks are selected from the THINC-it tool: 1. Spotter task, which is related to the attention function, lasts for 1 minute. There is a left arrow on the left and a right arrow on the right of the iPad. A yellow arrow appears consecutively between the two. This test requires participants to click the direction indicated by the yellow arrow as soon as possible. There are 20 yellow arrows in total. Finally, this test measures the reaction time of each time (millisecond), the number of correct answers, the maximum number of consecutive correct answers. The results are recorded on the iPad. Therefore, we can figure out mean of the log-transformed reaction time. 2. Symbol Check, which is related to the working memory function, lasts for 1 minute. On the bottom of the iPad, there are five fixed symbols, the center of the screen scroll presents a series of symbols. The Subjects is expected to click the “Same” button within several seconds. The first six symbols are displayed. The seventh begins to be hidden so that the participants must remember the next symbol. There were 20 attempts totally. Finally, we can calculate the correct number; 3. Code Break, which is related to memory, executive control, and attention function, lasts for 2 minutes. An iPad reads out a series of numbers corresponding to six symbols to the test participant, and participants are then required to click the matching symbol as quickly as possible. There were 40 attempts in two minutes. The correct number is calculated. Schematic diagrams of the tasks are shown in Fig. 1. The subjects only performed one sole trial of each test. There is a 1-minute rest between each task. The subjects are told to avoid any major body movement during the experiment. Before the experiment, a practice session is held to familiarize the subjects with the task instructions.
Fig. 1. (1) VFT: The participants were asked to list as many animals as they can in 60 s. (2) Spotter: The yellow arrows appear randomly on the screen. The participants should tap the direction key according to the yellow arrow as fast as they can, their reaction causes the next arrow to appear. (3) Symbol-Check: Five symbols are randomly arranged in the center of the screen and scroll. The participants are presented with a laterally moving sequence of symbols, the first of which is then hidden. The users must correctly recall the hidden symbol as quickly as possible. (4) Code-Break: There are six numbers corresponding to six symbols. The participants should tap the corresponding shape at the bottom of the screen according to the number as fast as they can. All the tasks were presented as a single trial for each subject.
2.3 fNIRS measurement
The changes in HbO and deoxygenated hemoglobin (Hb) were measured using the 52-channel ETG-4000 Optical Topography System (Hitachi Medical Corporation, Tokyo, Japan). The distance between each pair of detector probes was set to 3cm, and we defined each measuring area of each group of detector probes as a channel. A 3 × 11 channel probe set was placed on the participants’ foreheads so that the middle probe in the lowest row was placed at the Fpz position according to the international 10–20 System for electrode placement [24]. The channel locations in this study were shown in Fig. 2. The near-infrared light was emitted in two wavelengths (695 ± 20 nm and 830 ± 30 nm) by 17 laser diodes and the relative changes of the reflected light were measured by 16 photo detectors. The sampling frequency of the recording was set to 10 Hz. The fNIRS recording system can exclude the bad channels with low signal-to-noise ratio for each subject. The signal-to-noise ratios were calculated by the coefficient of variation of the recorded signals.
Fig. 2. (a) Locations of the near-infrared NIRS probe. The NIRS channel 16 was in the midline of standard brain space. (b) NIRS probe and channel setting. Red ovals represent near-infrared light emitter, blue ovals represent near-infrared light detectors, and lines represent NIRS channels. The locations of NIRS channels were estimated probabilistically according to the international 10-20 system.
2.4 Data analysis
A band-pass filter between 0.015 and 0.5 Hz was applied for the original data. The filter was least-squares finite impulse response (FIR) filter with zero-phase distortion and the filter order was 50 [17]. Modified Beer-Lambert law [25] was used to transform optical signals into hemodynamic parameters. A correlational-based signal improvement (CBSI) [26] was applied.
GLM is applied by using a gauss function as HRF (peak time = 6.0 s; Full Width Half Maximum (FWHM) = 5.89) and its first and second temporal derivative to modulate the onset and the dispersion of the HRF [27,34]. Estimated beta weights for corrected HbO measures were determined using the ordinary least squares regression analysis [28]. The beta estimators (also called estimated beta) were calculated by the time series GLM. The data matrix Y of order (TxC) containing the functional NIRS time series T of each channel C is predicted by X consisting of a set of reasonable hemodynamic response functions (HRFs) which are convolved with the event sequence (the order of X is (TxM) where M is the number of modeled effects—see below). The functional data can be modeled as:
where X is the design matrix and β is the parameter matrix. In the simplest case, each column M of matrix X contains the predicted hemodynamic response for one experimental condition over time (T). To address inter-individual differences regarding the HRF’s latency and dispersion, the inclusion of the HRF’s first and second temporal derivative has been proposed. The inclusion of derivative terms results in an extension of the design matrix X: for one experimental condition, X contains two (HRF+1. Derivative) or three columns (HRF+1. Derivative+2. Derivative).The ordinary least square estimates of β are given by
The β-weights quantify the contribution of a predictor (e. g. HRF) for explaining the functional time series Y and serve as the parameter set for subsequent hypothesis testing.
More detailed information about GLM and estimated beta can be found in [27]. We used the data of the entire stimulation period for GLM. The peak of the hemodynamic response was set to 7 s after the stimulus onset. We used the data of whole task period for GLM.
We collected three types of bipolar scores including the Hamilton Anxiety Scale (HAMA), HAMD, and Montgomery Depression Scale (MADRS) for each subject in BD-D. Pearson correlations were used to analyze the relationship between hemodynamic responses and multi clinical bipolar scores.
The wavelet transform coherence (WTC) method was used to calculate pairwise coherences between all channels of the same probe to obtain the intra-hemispheric functional connectivity for each side of the PFC [29]. Only HbO signals were analyzed because Hb is much more sensitive to the crosstalk than HbO2 and has low signal to noise ratio [30]. The frequency bands in which the coherence value changed between the rest and the memory task were identified as the task-related bands to remove interference unrelated to cognitive tasks [29]. The average coherence value for these frequency bands, which was between 0.071 and 0.2 Hz, was calculated to indicate functional connectivity for each channel pair. The time resolution of WTC is 0.1s and the frequency resolution is about 0.01 Hz. The WTC Matlab package downloaded from their website (http://noc.ac.uk/using-science/crosswavelet-wavelet-coherence) was employed to evaluate intrahemispheric functional connectivity for the left and the right PFCs separately.
For the statistical analysis and drawing results’ figures in this study, we used the Seaborn toolbox in Python, which is an open-source programming software. We used two-sample unpaired t-test of the null hypothesis of equal means with two sided tails. Before the t-tests in this study, the normal distribution of the data was tested using Lilliefors test. Normal distribution is a precondition of t-test. Level of significance was taken as p-values (p) <0.05
3. Results
3.1 Behavior performances
Performances of BD-D were poorer than HC group in all the four tasks (Fig. 3). Behavior performances of three tasks from Thinc-it showed significant difference between groups.
Fig. 3. Behavior results obtained during tasks. The data are expressed as mean values ± SE. The red fonts indicate p < 0.05. The degrees of freedom (DFs) were 48.
3.2 Hemodynamic responses
The channels were divided into left and right groups. We averaged hemodynamic changes of channels in left or right group for each subject (Fig. 4). Increase of HbO and decrease of Hb can be seen during the tasks, which represented brain activation during tasks in both left and right channels. The hemodynamic changes were lower in BD-D than in HC during the VFT, Symbol-Check and Code-Break tasks. To correct baseline, we subtract the 2s mean signal value before task beginning from the whole task series for each task.
Fig. 4. Grand average of original hemoglobin signal time series for the VFT and the Symbol-Check tasks. 0 s represents the beginning of task and 60 s represents the end of task. The grey region indicates standard error (SE).
Beta estimators of HbO were shown as violin and scatter diagrams (Fig. 5). HbO activation was stronger in HC than in BD-D during the VFT, Symbol-Check and Code-Break tasks. Lateralization difference only existed in the Symbol-Check task. We did not give the results of Hb because that the alternations of Hb are negative correlated with HbO after CBSI.
Fig. 5. Average activation of HbO in HC and BD-D groups. T-tests between groups were done. Significant differences (p < 0.05) were marked in red. The degrees of freedom (DFs) were 48.
The activation difference between groups is clearly in the mean HbO activation (estimated beta) maps (Fig. 6). The pseudo-color maps were made by cubic interpolation of HbO activation averaged across BD-D and HC. A two-dimensional matrix was formed to simulate the map of PFC area. These estimated beta values were filled into the matrix and their line and column numbers were correlated with their channel locations. For other position of the matrix, two-dimensional cubic interpolation was used to get these values referred the existed activation values. Activation in HC is significantly higher and broader than in BD-D during the VFT and the Symbol-Check tasks. Figure 7 shows the second-level results evoked by the four cognitive tasks overlaid on a standard brain.
Fig. 6. Pseudo-color maps constructed from fNIRS channels of HbO activity during four tasks using cubic interpolation. L represents the left cortex and R represents the right cortex. Color-bars were showed below for VFT and Symbol Check separately.
Fig. 7. T-maps comparing the prefrontal activation elicited by the four tasks for the HC and BD-D groups.
3.3 Functional connectivity
WTC values were shown as violin and scatter diagrams (Fig. 8). Significant differences between BD-D and HC only exist in the Symbol-Check and Code-Break tasks. Lateralization differences also existed in these two tasks.
Fig. 8. Average intra-hemispheric functional connectivity of the HbO signals. T-tests between groups were done. Significant differences (p < 0.05) were marked in red. Fisher’s z-transform was employed for the coherence values before the unpaired t-test.
3.4 Correlation between estimated beta and scores of psychiatric disorders
According to the signal quality, we selected 20 channels in PFC for the correlation analysis. The 20 selected channels are 16, 23, 24, 25, 26, 27, 29, 30, 33, 34, 35, 36, 40, 44, 45, 47, 48, 49, 50. Their locations can be found in Fig. 2 of the revised manuscript. These channels had high signal-to-noise ratio were not excluded by the system in more than 28 BD subjects. The correlation matrix is shown in Fig. 9. Most channels are negative correlated with the Symbol-Check task, which means that the higher the risk of disease, the lower brain activation during working memory task. Only three channels’ hemodynamic activations showed positive correlation with the attention related task. For all p-values (p) in this section, false discovery rates (FDR) correction was used to control the expected proportion of ‘discoveries’ that are false.
Fig. 9. Correlation matrix. Task 1 is VFT. Task 2 is Spotter. Task 3 is Symbol-Check. Task 4 is Code-Break. The yellow rectangles represent significant positive correlation (p < 0.05). The green rectangles represent significant negative correlation (p < 0.05). The rectangles represent no correlation (p > 0.05).
4. Discussion
The goal of the study was to investigate the brain dysfunction of bipolar depression during multi-cognitive tasks. The time responses of hemodynamic alternations during the 4 tasks show promised response, and proved that the experimental design is reasonable and THINC-it tool has clinical potential. Presented results support the initial hypothesis of lower PFC activity in BD-D as compared to the HC group, measured by changes in mean corrected beta estimators for the HbO responses. PFC has been proven to be closely related to multi-cognitive functions and affected by cognitive impairment in fNIRS studies [31,32]. Lower activity in the left dlPFC were associated with executive functioning. Significant differences are observed in the right PFC [33,34]. These results are in line with a previous study that revealed lower PFC activation in individuals at risk for psychosis [35].
Considering a lot of outpatient cases every day in China, the diagnosis of each patient is a heavy burden for doctors. Therefore, a shorter experimental time is very important for both doctors and patients in the psychiatric clinic. We used a single trial experiment in this study. The duration of each task is contained within 2 minutes. It must be admitted that the VFT may reflect reasoning ability and perceptual speed, only a fraction of executive function, which encompass various aspects, such as response inhibition, working memory, and cognitive flexibility [36]. Research with multiple tests is required to replicate current findings [37].
Brain dysfunction during Symbol-Check mainly exists in the right PFC. The differences in brain activity between groups are in line with the difference in brain connectivity. Therefore, the fNIRS signals of the right PFC during working memory task can help doctors to diagnose BD patients. The Code-Break task also showed great potential to distinguish between BD-D and HC. However, the task duration is longer than the Symbol-Check task and the correlation with scores of psychiatric disorders is not as good as the Symbol-Check task.
In recent years, some studies used fNIRS to investigate the prefrontal activation of BD-D [33,38]. Compared to these studies, we used experiments with much shorter durations. To our knowledge, this study is the first time to analyze the fNIRS based brain connectivity of BD-D. Although the results of brain connectivity are not as sensitive as brain activity in distinguishing groups, the brain connectivity can be used as a supplementary basis to reduce the misdiagnosis rate. We will investigate connectivity analysis more in-depth in our future studies.
Correlation between estimated beta and bipolar scores indicates that the Symbol Check task can better reflect the degree of brain dysfunction for the BD-D subjects. This result also supports that Symbol-Check has great potential in BD diagnosis. A previous study demonstrated that MADRS is as sensitive an instrument as HAM-D for detecting antidepressant efficacy in clinical trials [39]. Another study showed that the MADRS reliability statistics were higher than those of the HAM-D for detecting initial symptoms of unipolar depression [40].
Only three channels (channels 34, 44, and 45) in the Spotter task showed positive correlations with scores of psychiatric disorders. The three channels are all located in a triangular area of the right VLPFC (Fig. 10), indicating that BD-D subjects may show increased brain activity in right VLPFC during attention related tasks. We will pay more attention to this area in future study.
We will explore the methods based on fNIRS to distinguish regular depression from BD-D. For example, more features of fNIRS include brain connectivity, effective connectivity measure that distinguish these two states need to be extracted. And electroencephalogram can be used as a simultaneous detection method to provide a more comprehensive information of neuro activity. In addition, we will explore more experimental paradigms to distinguish the specific brain function differences between regular depression and BD-D.
5. Conclusions
In conclusion, this study suggests that the Symbol Check task may be valuable to assist the clinical diagnosis of BD. The Code Break task also has the potential to be used in clinical diagnosis. However, to confirm this conclusion, more experiments are needed with whole brain multimodal brain function tests. Meanwhile, we will explore and optimize a new experimental paradigm that is more suitable for the clinical diagnosis of psychological diseases based on this study.
Funding
National Natural Science Foundation of China (81971660); Chinese Academy of Medical Science health innovation project (2021-I2M-042, 2021-I2M-058); Sichuan Science and Technology Program (2021YFH0004); Tianjin Outstanding Youth Fund Project (20JCJQIC00230); Basic Research Program for Beijing-Tianjin-Hebei Coordinatio (19JCZDJC65500(Z)); Fundamental Research Funds for the Central Universities (No. 3332019101); Program of Chinese Institute for Brain Research in Beijing (2020-NKX-XM-14); Zhejiang Provincial Key Research and Development Program (2021C03107).
Acknowledgements
This study was supported by the National Natural Science Foundation of China (No. 81971660), Chinese Academy of Medical Science health innovation project (2021-I2M-042, 2021-I2M-058), Sichuan Science and Technology Program (No. 2021YFH0004), Tianjin Outstanding Youth Fund Project (No. 20JCJQIC00230), Program of Chinese Institute for Brain Research in Beijing(2020-NKX-XM-14), Basic Research Program for Beijing-Tianjin-Hebei Coordination (19JCZDJC65500(Z)), Fundamental Research Funds for the Central University (No. 3332019101) and Zhejiang Provincial Key Research and Development Program (No. 2021C03107).
Disclosures
The authors declare no conflicts of interest.
Authorship contribution statement Chenyang Gao: Writing original draft, Formal analysis. Hetong Zhou: Conceptualization, Formal analysis, Data curation. Jingjing Liu: Data curation. Jia Xiu: Formal analysis. Qi Huang: Data curation. Yin Liang: Data curation. Ting Li: Conceptualization, Writing - review & editing, Funding acquisition. Shaohua Hu: Conceptualization, Writing - review & editing, Funding acquisition.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. G. S. Malhi, D. M. Bargh, C. M. Coulston, P. Das, and M. Berk, “Predicting bipolar disorder on the basis of phenomenology: implications for prevention and early intervention,” Bipolar Disord. 16(5), 455–470 (2014). [CrossRef]
2. M. Bauer, O. A. Andreassen, J. R. Geddes, L. Vedel Kessing, U. Lewitzka, T. G. Schulze, and E. Vieta, “Areas of uncertainties and unmet needs in bipolar disorders: clinical and research perspectives,” Lancet Psychiatry. 5(11), 930–939 (2018). [CrossRef]
3. C. Bourne, Ö Aydemir, V. Balanzá-Martínez, E. Bora, S. Brissos, J. T. Cavanagh, L. Clark, Z. Cubukcuoglu, V. V. Dias, S. Dittmann, I. N. Ferrier, D. E. Fleck, S. Frangou, P. Gallagher, L. Jones, T. Kieseppä, A. Martínez-Aran, I. Melle, P. B. Moore, M. Mur, A. Pfennig, A. Raust, V. Senturk, C. Simonsen, D. J. Smith, D. S. Bio, M. G. Soeiro-de-Souza, S. D. Stoddart, K. Sundet, A. Szöke, J. M. Thompson, C. Torrent, T. Zalla, N. Craddock, O. A. Andreassen, M. Leboyer, E. Vieta, M. Bauer, P. D. Worhunsky, C. Tzagarakis, R. D. Rogers, J. R. Geddes, and G. M. Goodwin, “Neuropsychological testing of cognitive impairment in euthymic bipolar disorder: an individual patient data meta-analysis,” Acta Psychiatr Scand. 128(3), 149–162 (2013). [CrossRef]
4. M. M. Kurtz and R. T. Gerraty, “A meta-analytic investigation of neurocognitive deficits in bipolar illness: profile and effects of clinical state,” Neuropsychology 23(5), 551–562 (2009). [CrossRef]
5. L. J. Robinson, J. M. Thompson, P. Gallagher, U. Goswami, A. H. Young, I. N. Ferrier, and P. B. Moore, “A meta-analysis of cognitive deficits in euthymic patients with bipolar disorder,” J. Affective Disord. 93(1-3), 105–115 (2006). [CrossRef]
6. S. Koike, Y. Nishimura, R. Takizawa, N. Yahata, and K. Kasai, “Near-Infrared spectroscopy in schizophrenia: a possible biomarker for predicting clinical outcome and treatment response,” Front. Psychiatry 4, 145 (2013). [CrossRef]
7. K. Miller E and D. Cohen J, “An integrative theory of prefrontal cortex function,” Annu. Rev. Neurosci. 24(1), 167–202 (2001). [CrossRef]
8. P. Bebbington, “The world health report 2001,” Social Psychiatry and Psychiatric Epidemiology 36(10), 473–474 (2001). [CrossRef]
9. T Li, Y Lin, Y Shang, L He, C Huang, M Szabunio, and G. Yu, “Simultaneous measurement of deep tissue blood flow and oxygenation using noncontact diffuse correlation spectroscopy flow-oximeter,” Sci. Rep. 3(1), 1358 (2013). [CrossRef]
10. T. Li, C. Xue, P. Wang, Y. Li, and L. Wu, “Photon penetration depth in human brain for light monitoring and treatment: A realistic Monte Carlo Simulation study,” J. Innov. Opt. Health Sci. 10(05), 1743002 (2017). [CrossRef]
11. A. Villringer and B. Chance, “Non-invasive optical spectroscopy and imaging of human brain function,” Trends Neurosci. 20(10), 435–442 (1997). [CrossRef]
12. Y. Hoshi, “Functional near-infrared spectroscopy: current status and future prospects,” J. Biomed. Opt. 12(6), 062106 (2007). [CrossRef]
13. AC Ehlis, S Schneider, T Dresler, and AJ Fallgatter, “Application of functional near-infrared spectroscopy in psychiatry,” NeuroImage 85, 478–488 (2014). [CrossRef]
14. AV Medvedev, JM Kainerstorfer, SV Borisov, and VanMeter, “Functional connectivity in the prefrontal cortex measured by near-infrared spectroscopy during ultrarapid object recognition,” J. Biomed. Opt. 16(1), 016008 (2011). [CrossRef]
15. C. Mesquita R, A. Franceschini M, and A. Boas D, “Resting state functional connectivity of the whole head with near-infrared spectroscopy,” Biomed. Opt. Express 1(1), 324–336 (2010). [CrossRef]
16. M. F. Kaminski, S. Thomas-Gibson, M. Bugajski, M. Bretthauer, C. J. Rees, E. Dekker, G. Hoff, R. Jover, S. Suchanek, M. Ferlitsch, J. Anderson, T. Roesch, R. Hultcranz, I. Racz, E. J. Kuipers, K. Garborg, J. E. East, M. Rupinski, B. Seip, C. Bennett, C. Senore, S. Minozzi, R. Bisschops, D. Domagk, R. Valori, C. Spada, C. Hassan, M. Dinis-Ribeiro, and M. D. Rutter, “Performance measures for lower gastrointestinal endoscopy: a European Society of Gastrointestinal Endoscopy (ESGE) quality improvement initiative,” Endoscopy 49(04), 378–397 (2017). [CrossRef]
17. C. Gao, J. Sun, X. Yang, and H. Gong, “Gender differences in brain networks during verbal Sternberg tasks: A simultaneous near-infrared spectroscopy and electro-encephalography study,” J. Biophotonics 11(3), e201700120 (2018). [CrossRef]
18. M. A. Rahman, A. B. Siddik, T. K. Ghosh, F. Khanam, and M. Ahmad, “A narrative review on clinical applications of fNIRS,” Journal of Digital Imaging 33(5), 1167–1184 (2020). [CrossRef]
19. Y. Wei, Q. Chen, A. Curtin, L. Tu, X. Tang, Y. Tang, L. Xu, Z. Qian, J. Zhou, C. Zhu, T. Zhang, and J. Wang, “Functional near-infrared spectroscopy (fNIRS) as a tool to assist the diagnosis of major psychiatric disorders in a Chinese population,” European archives of psychiatry and clinical neuroscience 271(4), 745–757 (2021). [CrossRef]
20. R Takizawa, M Fukuda, S Kawasaki, K Kasai, M Mimura, S Pu, T Noda, S Niwa, and Y Okazaki, “Neuroimaging-aided differential diagnosis of the depressive state,” NeuroImage 85(1), 498–507 (2014). [CrossRef]
21. S. G. Costafreda, C. H. Fu, M. Picchioni, T. Toulopoulou, C. McDonald, E. Kravariti, M. Walshe, D. Prata, R. M. Murray, and P. K. McGuire, “Pattern of neural responses to verbal fluency shows diagnostic specificity for schizophrenia and bipolar disorder,” BMC Psychiatry 11(1), 18 (2011). [CrossRef]
22. D. Raucher-Chéné, A. M. Achim, A. Kaladjian, and C. Besche-Richard, “Verbal fluency in bipolar disorders: A systematic review and meta-analysis,” J. Affective Disord. 207, 359–366 (2017). [CrossRef]
23. S. H. Ho C, W. B. Zhang M, and R. Ho, “Optical topography in psychiatry: a chip off the old block or a new look beyond the mind–brain frontiers?” Front. Psychiatry 7, 74 (2016). [CrossRef]
24. H. Jasper, “Report of the committee on methods of clinical examination in electroencephalography,” Electroencephalogr. Clin. Neurophysiol. 10(2), 370–375 (1958). [CrossRef]
25. D. T. Delpy, M. Cope, P. van der Zee, S. Arridge, S. Wray, and J. Wyatt, “Estimation of optical pathlength through tissue from direct time of flight measurement,” Phys. Med. Biol. 33(12), 1433–1442 (1988). [CrossRef]
26. X. Cui, S. Bray, and A. L. Reiss, “Functional near infrared spectroscopy (NIRS) signal improvement based on negative correlation between oxygenated and deoxygenated hemoglobin dynamics,” NeuroImage 49(4), 3039–3046 (2010). [CrossRef]
27. M. M. Plichta, S. Heinzel, A. C. Ehlis, P. Pauli, and A. Fallgatter, “Model-based analysis of rapid event-related functional near-infrared spectroscopy (NIRS) data: a parametric validation study,” NeuroImage 35(2), 625–634 (2007). [CrossRef]
28. S. V. Tupak, A. Reif, P. Pauli, T. Dresler, M. J. Herrmann, K. Domschke, C. Jochum, E. Haas, C. Baumann, H. Weber, A. J. Fallgatter, J. Deckert, and A.-C. Ehlis, “Neuropeptide S receptor gene: fear-specific modulations of prefrontal activation,” NeuroImage 66, 353–360 (2013). [CrossRef]
29. A. Grinsted, C. Moore J, and S. Jevrejeva, “Application of the cross wavelet transform and wavelet coherence to geophysical time series,” Nonlin. Processes Geophys. 11(5/6), 561–566 (2004). [CrossRef]
30. J. MacIntosh B, M. Klassen L, and S. Menon R, “Transient hemodynamics during a breath hold challenge in a two part functional imaging study with simultaneous near-infrared spectroscopy in adult humans,” NeuroImage 20(2), 1246–1252 (2003). [CrossRef]
31. F. S. Racz, P. Mukli, Z. Nagy, and A. Eke, “Increased prefrontal cortex connectivity during cognitive challenge assessed by fNIRS imaging,” Biomed. Opt. Express 8(8), 3842–3855 (2017). [CrossRef]
32. D. Yang, K. S. Hong, S. H. Yoo, and C. S. Kim, “Evaluation of neural degeneration biomarkers in the prefrontal cortex for early identification of patients with mild cognitive impairment: an fNIRS study,” Front. Hum. Neurosci. 13, 317 (2019). [CrossRef]
33. L. Fu, D. Xiang, J. Xiao, L. Yao, Y. Wang, L. Xiao, and Z. Liu, “Reduced prefrontal activation during the Tower of London and verbal fluency task in patients with bipolar depression: a multi-channel NIRS study,” Front. Psychiatry 9, 214 (2018).
34. C. Gui-Fang, M. Meng-Chai, and F. Kun, “Brain activation during verbal fluency task in type II bipolar disorder patients: a near-infrared spectroscopy study,” Psychiatry Research 298, 113762 (2021). [CrossRef]
35. A. Aleksandrowicz, F. Hagenmuller, H. Haker, K. Heekeren, A. Theodoridou, S. Walitza, A. C. Ehlis, A. Fallgatter, W. Rössler, and W. Kawohl, “Frontal brain activity in individuals at risk for schizophrenic psychosis and bipolar disorder during the emotional Stroop task–an fNIRS study,” NeuroImage: Clinical 26, 102232 (2020). [CrossRef]
36. B. Jurado M and M. Rosselli, “The elusive nature of executive functions: a review of our current understanding,” Neuropsychol Rev 17(3), 213–233 (2007). [CrossRef]
37. Y. Liang, X. Jiang, W. Zhu, Y. Shen, F. Xue, Y. Li, and Z. Chen, “Disturbances of dynamic function in patients with bipolar disorder I and its relationship with executive-function deficit,” Front. Psychiatry 11, 537981 (2020). [CrossRef] .
38. Y. Zhu, W. Quan, H. Wang, Y. Ma, J. Yan, H. Zhang, W. Dong, and X. Yu, “Prefrontal activation during a working memory task differs between patients with unipolar and bipolar depression: A preliminary exploratory study,” J. Affective Disord. 225, 64–70 (2018). [CrossRef]
39. A. Khan, S. R. Khan, E. B. Shankles, and N. L. Polissar, “Relative sensitivity of the Montgomery-Asberg Depression Rating Scale, the Hamilton Depression rating scale and the Clinical Global Impressions rating scale in antidepressant clinical trials,” International Clinical Psychopharmacology 17(6), 281–285 (2002). [CrossRef]
40. M. Carneiro A, F. Fernandes, and A. Moreno R, “Hamilton depression rating scale and montgomery-asberg depression rating scale in depressed and bipolar I patients: psychometric properties in a Brazilian sample,” Health Qual Life Outcomes 13(1), 42 (2015). [CrossRef]
