| The rhythmicity, as the essential characteristic of neural activity, refers to the synchronization of a cluster of neurons in a particular frequency. Specifically, neurons communicate with each other via phase synchronization at the same frequency. According to the latest synchronized gating hypothesis of neural communication, the synchronization of high frequency neural oscillations is gated by the phase of low frequency neural oscillations. The cross frequency phase-amplitude coupling is one of the important rules of neural oscillations. In addition, different brain regions occupy distinctive natural frequencies with higher-order regions working in lower frequencies. Functionally, neural oscillations in lower frequency are responsible for functional integration, whereas those in higher frequency are associated with local information processing. So far, considerable data from electrophysiological studies have revealed the essential role of high frequency neural oscillations in cognitive processing; however, the mechanisms of low frequency neural oscillations are largely unknown due to the limitation of electrophysiological approaches. Therefore, a huge gap exists in establishing an integrative theory of neural oscillations. Recently, the rapid development of functional magnetic resonance imaging(fMRI) technique provides a powerful tool to investigate low frequency neural oscillations because this technique has low frequency signal, high spatial resolution and is closely related to neural activities.By means of the fMRI technique, the current work dedicated to investigate 1) frequency characteristics of low frequency neural oscillations during resting-state and 2) low frequency steady-state brain response(lfSSBR) along with particular cognitive processing. Two aspects of this dissertation have been put forward:The first part of this work explored frequency characteristics of low frequency blood oxygen level dependent(BOLD) fluctuations. It has been suggested that BOLD fluctuations are frequency specific in different brain areas; however, some problems in investigating BOLD fluctuations are not well resolved such as the definition of frequency bands and the confounding effect of physiological noises. In a large sample(n=64) with physiological noises being regressed out, the present dissertation investigated frequency characteristics of BOLD fluctuations in a relative wide frequency range(0.01-0.25 Hz) using the sliding window method(window width: 0.03 Hz; step length: 0.01 Hz). We observed that BOLD fluctuations in the full frequency band show a sandwich-like distribution with higher amplitude in the middle of the brain and lower amplitude in the superior and inferior of the brain. The frequency characteristics of BOLD fluctuations manifest a cortical-limbic dichotomy with cortical regions show relatively higher amplitude in the lower frequency end, whereas the limbic system and subcortical structures show relatively higher amplitude in the higher frequency end. In a part of these regions, the frequency characteristics of BOLD fluctuations have extremely high inter-subject stability, indicating some common states of brain activities during resting-state. On the contrary, the frequency characteristics in most areas have low inter-subject stability, indicating that BOLD fluctuations in these regions are independent of real-time cognitive activities.In the second part, the concept of lfSSBR was first put forward; the modulation of low frequency BOLD fluctuations to brain functions was then investigated by means of lfSSBR. By imitating the idea of steady-state evoked potential(SSEP), the current dissertation explored low frequency neural oscillations along with cognitive processes, benefiting from the advantage of fMRI in spatial resolution and advantages of steady-state brain response such as stable amplitude across a long time, high signal-to-noise ratio and phase locking.First, the lfSSBR was successfully evoked by a simple reaction time task and a semantic comprehension task. The lfSSBR had similar waveform to the SSEP; meanwhile, it could capture lower frequency and locate more accurately than the SSEP. Brain regions in which the lfSSBR was evoked were largely overlapped with those activated by cognitive tasks; however, the mechanisms of lfSSBR and activation are different: 1) the lfSSBR measures the variability of BOLD signal whereas activation measures the mean value of that; 2) the lfSSBR is independent of neurovascular coupling whereas activation is dependent of that. In addition, the lfSSBR showed a sensorimotor bias, indicating the relationship between low frequency oscillations and neural flexibility. Therefore, lfSSBR is a new index in exploring cognitive-related low frequency brain response that is different from SSEP and brain activation.Second, we studied how lfSSBR modulates large scale brain networks based on the revised attention network test(ANT) which has high reliability. The alerting, orienting, and executive control networks were highly similar to each other in the meaning of measurement, therefore providing a suitable test-retest paradigm for studying lfSSBR. Nevertheless, attention network scores assessed by the classical ANT have very low reliability. The current dissertation obtained high reliable and accurate attention network scores, and revealed mutually inhibited inter-network relationship by combining the non-orthogonal contrast approach with block design. These results are of importance for the establishment of the theory of attention system and for the application of ANT in clinical and developmental psychology. Based on the revised ANT, we investigated the modulation of lfSSBR to large scale brain networks. We found that the lfSSBR modulated the functional connectivity of brain networks in a frequency specific way. Furthermore, the modulation of lfSSBR to gross brain excitability overwhelmed its modulation to different attention states. These findings provided strong evidence for the cortical excitability hypothesis of low frequency neural oscillations, benefitting for the improvement of the theory of neural oscillations.Last, we investigated how lfSSBR modulates brain networks in a face recognition task. The waveform of lfSSBR at the fundamental frequency was approximate to the sinusoid when the task was shown at 0.05 Hz. At the ascending and descending phases rather than the peak and trough of lfSSBR, the functional connectivity in the right fusiform face area was significantly modulated, indicating that the modulation of lfSSBR to brain network was phase dependent. This finding emphasized the essential role of the phase of low frequency neural oscillations in cognitive functions, supporting the synchronized gating hypothesis.In summary, the current dissertation has two principle contributions: first, we revealed that low frequency BOLD fluctuations were constrained by functional and anatomical organization of the brain. A sandwich-like spatial distribution was found for the amplitude and a cortical-limbic dichotomy was observed for frequency characteristics; second, we innovatively put forward the index of lfSSBR to study low frequency neural oscillations during specific cognitive activities. Using this index, we uncovered some important rules of low frequency oscillations, supporting the latest assumptions of low frequency neural oscillations such as the cortical excitability hypothesis and the synchronized gating hypothesis. These findings contributed greatly to the establishment of an integrative theory of neural oscillations. |
PDF Full Text Request