Contemporary Consciousness Research: Methods and Experiments

Philosophical Framework

Contemporary consciousness research, despite its empirical orientation, is deeply rooted in philosophical traditions that pose fundamental questions and dilemmas. Central here is the so-called “hard problem of consciousness,” formulated by David Chalmers. It contrasts with the “easy problems,” which concern explaining cognitive functions such as attention, memory, or sensory information processing. The hard problem asks why and how physical processes in the brain give rise to subjective experience, or qualia — what it is like to be in a particular mental state [Block, 1995]. This question goes beyond purely functional explanations and requires understanding the phenomenal aspect of consciousness.

Empirical studies aimed at identifying neural correlates of consciousness (NCC) inevitably face this philosophical dichotomy. They attempt to establish which brain processes correspond to conscious experience but do not provide a direct answer to the nature of that experience. Philosophical arguments, such as the thought experiments involving “zombies” or “Mary’s room,” continue to stimulate discussions and guide empirical research, highlighting the gap between objective brain descriptions and subjective experience. This ongoing interaction between philosophy and science shapes the interdisciplinary landscape of consciousness research, where each field contributes to understanding this complex phenomenon.

Introduction

Consciousness research has undergone significant changes over recent decades, shifting from predominantly philosophical speculation to active empirical work employing advanced neuroscience methods. Today, consciousness is viewed as a multidimensional phenomenon requiring an interdisciplinary approach that integrates neuroscience, psychology, philosophy, and informatics [Overgaard, 2017]. A key challenge remains understanding how physical brain processes generate subjective experience, often referred to as the “hard problem of consciousness.” Contemporary research seeks to identify neural correlates of consciousness (NCC), the minimal set of neural mechanisms sufficient for conscious experience.

Various experimental paradigms, such as binocular rivalry and masking, are actively employed to manipulate conscious perception of stimuli while keeping sensory input constant. These methods help reveal differences in brain activity between conscious and unconscious states [Klink et al., 2015]. Concurrently, theoretical models like Integrated Information Theory (IIT) [Tononi, 2004] and Global Neuronal Workspace Theory (GNWT) offer explanations of consciousness architecture and functions and attempt to quantitatively assess its level. These models not only explain existing data but also generate new hypotheses for empirical testing, fostering further progress in this complex field.

Research Overview

Diagnosis and Assessment of Disorders of Consciousness

How can we be sure that a person lying silently and motionless is truly unconscious? This seemingly simple question lies at the heart of one of the most challenging problems in modern medicine and neuroscience: accurate diagnosis and assessment of disorders of consciousness. For a long time, clinical evaluation based on behavioral responses was the only available tool. However, research has shown that this approach often leads to misdiagnoses. For example, Schnakers and colleagues in 2009 noted that up to 43% of patients with disorders of consciousness could be mistakenly diagnosed as being in a vegetative state (VS), while in reality, they retained minimal consciousness. This is not merely a statistical error; it is a tragedy for the patient and their family, as an incorrect diagnosis may deprive the person of a chance for rehabilitation and appropriate treatment.

The problem lies in the fact that behavioral manifestations of consciousness can be extremely sparse or entirely absent even when internal awareness exists. A patient may be “locked in” their own body, unable to move or speak but still capable of perceiving and even thinking. Here, neuroimaging methods come to the rescue, promising to revolutionize diagnosis. Wang et al. (2023) emphasize that neuroimaging methods can detect conscious activity in patients who show no behavioral signs of consciousness and provide objective and quantitative indicators to assist clinicians in diagnosis [Wang et al., 2023]. This opens entirely new prospects for detecting hidden consciousness.

One of the most promising directions is the use of functional magnetic resonance imaging (fMRI) and positron emission tomography (PET). These methods measure brain metabolic activity and blood flow closely linked to neural activity. For example, studies using 18F-FDG PET/CT, as shown by Zhao et al. (2018), allow assessment of brain function in patients with disorders of consciousness by identifying glucose metabolism patterns that may indicate residual conscious activity. Similarly, Yamaki et al. (2023) used resting-state glucose metabolism to predict voluntary upper limb movements in patients with chronic severe brain injury, demonstrating the potential of these methods for evaluating functional recovery.

Electroencephalography (EEG) also plays a vital role, especially due to its high temporal resolution. It records electrical brain activity and identifies specific patterns associated with information processing and consciousness. For example, Zhang et al. (2022) showed that analyzing brain network properties in EEG during task performance can help assess residual motor function in patients with disorders of consciousness. Moreover, Wu et al. (2018) investigated the effect of auditory stimuli on patients with disorders of consciousness using quantitative EEG, finding changes in brain electrical activity that may indicate responses to external stimuli. These studies demonstrate that even in the absence of behavioral reactions, the brain can actively respond to the environment.

However, despite significant progress, none of these methods is a panacea. Each has its limitations. fMRI, for instance, offers excellent spatial resolution but relatively low temporal resolution, complicating tracking rapid changes in consciousness. EEG, conversely, has high temporal but low spatial resolution, making precise localization of activity sources challenging. Therefore, as Wang et al. (2023) note, single-modality neuroimaging no longer meets researchers’ needs [Wang et al., 2023]. Multimodal approaches combining several methods, such as simultaneous EEG-fMRI, provide a more comprehensive picture, compensating for the shortcomings of each individual method.

Besides neuroimaging, standardized behavioral scales like the Coma Recovery Scale-Revised (CRS-R), developed by Giacino et al. (2002), remain important tools. Although prone to errors, their systematic application and staff training can significantly improve diagnostic accuracy. These scales assess various aspects of consciousness, including auditory, visual, motor, oropharyngeal, communicative, and general functions, helping differentiate vegetative state from minimally conscious state [Giacino et al., 2002]. Schnakers et al. (2009) emphasize that the minimally conscious state (MCS) is characterized by inconsistent but clearly discernible behavioral signs of consciousness, requiring careful and repeated assessment to detect these subtle manifestations [Schnakers et al., 2009].

It is important to note that even with the most advanced neuroimaging methods, data interpretation remains challenging. Brain activity detected by fMRI or EEG does not always unequivocally indicate the presence of consciousness. For example, the brain may respond to stimuli at a subconscious level without forming subjective experience. This raises the fundamental question of what exactly we measure when we talk about “conscious activity.” David Chalmers called this the “hard problem of consciousness” — explaining why and how physical brain processes generate subjective experience (qualia). Neuroimaging can help identify neural correlates of consciousness (NCC) but does not explain the nature of subjective experience itself.

Nevertheless, neuroimaging provides objective data that can be used to refine diagnosis and predict outcomes. For example, studies have shown that patients demonstrating brain activation in response to instructions, even without behavioral reactions, have better recovery prognosis. Moreover, some studies, such as Kempny et al. (2018), showed that patients with prolonged disorders of consciousness may exhibit classic EEG responses to their own name compared to others’ names, indicating preserved ability to process meaningful information. Diagnosis of disorders of consciousness is not a static process but a dynamic field where clinical assessment and neuroimaging methods continuously improve and complement each other. Abandoning outdated concepts and actively implementing new technologies allows not only improving diagnostic accuracy but also potentially detecting hidden consciousness where it was previously unseen. However, despite all achievements, the question remains open: how can these diagnostic tools be integrated into effective rehabilitation strategies to not only detect consciousness but also help restore it?

Methods of Rehabilitation for Disorders of Consciousness

Having discussed the complexities of diagnosis and assessment of disorders of consciousness, it is logical to move to the question of how we can intervene and help patients in these states. Accurate state determination is only the first step; the true goal of medicine and neuroscience is to find ways to restore or improve consciousness. This is not merely a medical task but a deeply ethical one, as it touches on the very essence of human existence and the capacity to interact with the world.

One of the most promising, though invasive, approaches is deep brain stimulation (DBS). As Yang et al. note, among existing research on the treatment of disorders of consciousness (DOC), DBS offers a highly promising therapeutic approach. This method involves surgical implantation of electrodes into specific brain areas, which are then stimulated electrically. The goal of DBS is to modulate neural network activity believed to play a key role in maintaining consciousness. For example, stimulation of the central thalamus may promote restoration of connections between different cortical areas, critical for information integration and conscious experience emergence. However, despite promising results, DBS remains a complex and costly procedure requiring careful patient selection and risk assessment.

Besides invasive methods, non-invasive neuromodulation approaches are actively developing. A meta-analysis by Liu et al. demonstrated the effectiveness of methods such as transcranial direct current stimulation (tDCS), transcranial magnetic stimulation (TMS), and median nerve stimulation. These methods modulate brain activity without surgery, making them more accessible and safer. tDCS uses weak electrical currents to alter neuron excitability in specific cortical areas, while TMS generates magnetic fields inducing electrical currents in the brain. Effectiveness is typically evaluated by changes in behavioral responses or neurophysiological indicators, such as EEG patterns, which may signal increased consciousness levels. However, as with DBS, precise mechanisms and optimal stimulation parameters remain under active investigation.

Interestingly, even relatively simple methods like auditory stimulation can positively affect patients with disorders of consciousness. Studies by Zhu et al. and Wu et al. showed that presenting familiar music or voices of loved ones can induce changes in brain activity and even improve behavioral responses. For instance, Boltzmann et al. found that auditory stimulation modulates resting-state functional connectivity in patients with unresponsive wakefulness syndrome. This underscores the importance of sensory stimulation and its potential to activate residual neural networks possibly related to consciousness. The question remains, however, to what extent these changes represent true restoration of consciousness rather than mere automatic stimulus responses.

Neuroimaging methods are crucial for evaluating the effectiveness of these rehabilitation approaches. Wang et al. emphasize that neuroimaging methods have proven effective in consciousness rehabilitation evaluation. They allow objective measurement of changes in brain activity and structure that may be imperceptible through behavioral observation. For example, PET reflects metabolic activity of brain cells, an important indicator of their functional state [Wang et al., 2023]. However, PET has limitations, including radioactive exposure, restricting its frequency of use.

In this context, EEG and fMRI become primary tools. EEG’s high temporal resolution enables tracking rapid changes in brain electrical activity related to information processing and consciousness emergence. Wang et al. note that EEG has made great progress in evaluation and classification of DoC patients in task-based paradigms due to its high temporal resolution. Conversely, fMRI offers high spatial resolution, allowing localization of active brain areas and study of functional connectivity.

Combining various neuroimaging methods, so-called multimodal approaches, promises even greater accuracy. Wang et al. indicate that multimodal methods can achieve complementary advantages of different methods, thereby uncovering significant relationships that cannot be detected by employing a single modality alone. For example, PET/MRI combination allows simultaneous assessment of brain structure and metabolism. Similarly, joint use of EEG and fMRI enables precise localization of brain activity in time and space, compensating for individual method limitations. This is especially important for detecting “hidden consciousness” in patients who show no external behavioral signs of awareness [Wang et al., 2023].

However, despite all technological advances, a fundamental question remains: what exactly are we rehabilitating? If we speak of restoring consciousness, should we understand it as the ability to respond to external stimuli or as an internal subjective experience? Velmans raises the problem of the “hard problem of consciousness,” pointing out that functional explanations cannot fully capture the phenomenal aspect of consciousness. Even if a patient begins to show behavioral reactions or their brain exhibits patterns similar to healthy individuals, this does not guarantee the presence of subjective experience.

Methods of rehabilitation for disorders of consciousness are rapidly evolving, offering both invasive and non-invasive approaches supported by increasingly sophisticated neuroimaging techniques. However, each method faces the challenge of result interpretation: are we truly restoring consciousness in its phenomenal sense, or merely improving functional responses? This question directly leads us to the need for a deeper understanding of the theoretical foundations of consciousness and its measurement, which will be the subject of our next discussion.

Theoretical Foundations of Consciousness and Its Measurement

After discussing rehabilitation methods for disorders of consciousness, a natural question arises: what exactly are we trying to rehabilitate? What is the theoretical basis of consciousness, and how can we measure it, especially when behavioral manifestations are absent? This is not merely an academic interest; understanding the nature of consciousness directly impacts the development of effective therapeutic strategies and diagnostic tools.

One of the most influential contemporary theories is Integrated Information Theory (IIT), developed by Giulio Tononi. According to IIT, consciousness is the capacity of a system to integrate information [Tononi, 2004]. This theory relies on two key phenomenological properties of consciousness: differentiation, i.e., the presence of a vast number of distinct conscious experiences, and integration, i.e., the unity of each such experience. In other words, conscious experience is simultaneously rich and unified. IIT proposes a quantitative measure of consciousness, called Φ (phi), reflecting the degree to which a system is unified and irreducible, meaning it cannot be decomposed into independent parts without loss of information. The higher the Φ, the higher the level of consciousness. This concept allows not only theorizing about consciousness but also seeking ways to measure it, which is critically important for patients with disorders of consciousness.

However, not everyone agrees with a unitary approach to consciousness. Philosopher Ned Block, for example, proposes distinguishing different kinds of consciousness, calling it a “mixed concept” [Block, 1995]. He differentiates phenomenal consciousness (P-consciousness), which is pure experience — what it is like to be in that state — and access consciousness (A-consciousness), characterized by information availability for reasoning, speech, and goal-directed action. For Block, phenomenal consciousness is the subjective quality of experience, such as the sensation of red or pain, whereas access consciousness is a cognitive function enabling us to use that information. This distinction has profound clinical implications: a patient may possess phenomenal consciousness but lack the ability to express it due to impaired access consciousness. How then can we measure P-consciousness if it does not manifest behaviorally?

In response to such challenges, Francisco Varela proposed a methodological approach called neurophenomenology [Varela, 1996]. This approach aims to establish mutual constraints between phenomena present in subjective experience and correlational fields studied by cognitive sciences. Neurophenomenology calls for systematic investigation of subjective experience (phenomenology) combined with objective neurobiological data. This means we cannot ignore the patient’s inner world even if they cannot express it. Instead of relying solely on external behavioral markers, neurophenomenology suggests seeking “articulations through mutual constraints” between what is experienced and what is measured in the brain.

Contemporary research actively uses neuroimaging methods to assess consciousness, especially in patients with disorders of consciousness (DoC). For example, functional near-infrared spectroscopy (fNIRS) allows non-invasive evaluation of brain activity. Studies have shown that fNIRS can be used to assess brain function in minimally conscious state (MCS) patients using motor imagery tasks [Wang et al., 2023]. When patients were asked to imagine movements, their hemodynamic responses resembled those of control groups. This confirms the possibility of using fNIRS to detect hidden consciousness. Moreover, resting-state fNIRS studies have shown that MCS patients and unresponsive wakefulness syndrome (UWS) patients exhibit different topological architecture and connectivity patterns in the prefrontal cortex, enabling differentiation of these states [Liu et al., 2023].

Positron emission tomography (PET) also plays an important role in consciousness assessment, especially in studying brain glucose metabolism. Studies using 18F-FDG PET/CT have shown that glucose metabolism levels in various brain regions differ significantly among UWS, MCS, and recovered consciousness patients. These data confirm the correlation between cerebral glucose metabolic rate and consciousness level. However, PET has limitations, including radiation exposure and low spatial resolution, restricting its widespread use [Wang et al., 2023].

Another approach to understanding consciousness is related to the concept of predictive coding, especially in the context of interoception. Anil Seth and colleagues proposed a model in which conscious presence and its disorders are explained by interoceptive prediction errors [Seth et al., 2012]. In this model, consciousness arises from the brain’s ability to predict and explain internal bodily signals (interoception). If brain predictions successfully suppress informative interoceptive signals, a sense of “presence” emerges. Disruptions in this process may lead to states such as depersonalization or derealization. This model links consciousness with the sense of one’s own body and its states, a fundamental aspect of subjective experience.

Interestingly, studies of altered states of consciousness, such as those induced by psychedelics, also provide clues about the nature of consciousness. Michael Schartner and colleagues found that psychedelic substances like psilocybin, ketamine, and LSD increase neural signal diversity, measured by entropy and Lempel-Ziv complexity [Schartner et al., 2017]. These measures are typically higher in wakefulness compared to reduced consciousness states like anesthesia. The fact that psychedelic states show even greater signal diversity suggests they may represent an “enhanced level of consciousness,” challenging traditional views of normal wakefulness as the peak of conscious experience.

In the context of clinical interventions, deep brain stimulation (DBS) also provides valuable data for understanding theoretical foundations of consciousness. Stimulation of the central thalamus, for example, has shown behavioral improvements in patients with severe traumatic brain injury. The thalamus, as a key node in arousal and attention circuits, plays a critical role in maintaining consciousness [Mashour et al., 2020]. Studies show that DBS can enhance EEG functional connectivity in minimally conscious state patients. These results not only offer new therapeutic possibilities but also support the hypothesis that consciousness depends on integration of activity in distributed neural networks, especially those involving the thalamus and prefrontal cortex.

Theoretical foundations of consciousness encompass a wide range of ideas, from integrated information to different types of consciousness and predictive coding. Measuring consciousness, especially under its impairment, requires a comprehensive approach combining neuroimaging, electrophysiology, and phenomenological methods. However, despite significant progress, we still face the fundamental question: how exactly do these theoretical constructs manifest at the level of neural activity? This leads us to the need for more detailed study of neural correlates of consciousness.

Neural Correlates of Consciousness (NCC) and Their Measurement

Having discussed theoretical foundations of consciousness and various measurement methods, a natural question arises: how do these abstract concepts manifest in brain function? Where exactly in neural activity can we find the imprint of conscious experience? This question is addressed by research on neural correlates of consciousness (NCC) — the minimal set of neural events and mechanisms sufficient for a specific conscious experience. Essentially, we seek brain activity patterns that always occur when conscious experience is present and always are absent when it is not.

The traditional approach to searching for NCC, dominant for decades, is the so-called contrastive analysis [Sandberg et al., 2014]. The idea is simple: compare brain activity in two conditions, one involving conscious perception of a stimulus and the other not, while keeping all other parameters as similar as possible. For example, present the same stimulus, but in one case it is consciously perceived, and in the other remains unconscious due to masking or binocular rivalry. The difference in neural activity between these conditions should indicate NCC. However, as Christian Sandberg and colleagues note, this method has limitations and potential pitfalls. It may conflate true NCC with precursors or consequences of consciousness, which are not part of the conscious experience itself but closely related [Sandberg et al., 2014].

To overcome these limitations, researchers have begun seeking more refined methods. One such approach is multivariate decoding of neural activity. Sandberg et al. emphasize that multivariate decoding allows identifying which neural activity patterns consistently predict conscious experience at the single-trial level [Sandberg et al., 2014]. This is critical because “true” NCC should be consistently predictable, whereas consequences of consciousness may not occur in every trial, and precursors may be present without conscious experience.

Multivariate analysis enables separating the wheat from the chaff by focusing on patterns truly integral to consciousness.

One promising direction in NCC measurement is the Perturbational Complexity Index (PCI), proposed by Giulio Tononi and colleagues [Casali et al., 2013]. PCI is based on Integrated Information Theory (IIT), which postulates that consciousness arises from a system’s capacity to integrate information. PCI is measured by perturbing the cerebral cortex using transcranial magnetic stimulation (TMS) and analyzing the spatiotemporal pattern of electrocortical responses. The more complex and diverse this response, the higher the level of integrated information and, consequently, consciousness. This method provides an objective, quantitative index of consciousness level, especially valuable clinically for diagnosing disorders of consciousness such as vegetative state or minimally conscious state [Schnakers et al., 2009].

Interestingly, changes in signal diversity are also observed in altered states of consciousness. Michael Schartner and colleagues found increased spontaneous signal diversity in MEG during psychoactive doses of ketamine, LSD, and psilocybin [Schartner et al., 2017]. They note that this increase is most pronounced for Lempel-Ziv complexity, which reflects temporal rather than spatial signal diversity. This indicates that psychedelic states are characterized not merely by altered activity but by increased complexity and unpredictability of neural patterns. These changes were most prominent in occipito-temporal areas, consistent with previous data on alpha-band alterations [Schartner et al., 2017].

Integrated Information Theory, developed by Giulio Tononi, offers a profound philosophical and mathematical foundation for understanding NCC. According to this theory, consciousness is not merely the presence of activity in certain brain areas but the system’s ability to generate a large amount of integrated information [Tononi, 2004]. Tononi argues that consciousness arises from the system’s capacity to be simultaneously differentiated (able to take many distinct states) and integrated (these states cannot be reduced to independent parts). He even suggests that activity in certain cortical circuits may not contribute to consciousness if these circuits implement informationally isolated loops outside the main thalamocortical complex [Tononi, 2004]. This raises the important question of which neural structures are necessary for consciousness and which are merely accompanying.

Another important theoretical framework is the Global Neuronal Workspace Hypothesis (GNWH), actively developed by Stanislas Dehaene and Jean-Pierre Changeux [Mashour et al., 2020]. According to this hypothesis, consciousness arises when information processed by specialized brain modules becomes available to a wide range of other modules through a global workspace. This workspace is a network of neurons with long-range connections capable of broadcasting information throughout the brain. NCC within GNWH is not a localized area but rather a dynamic state in which information is “broadcast” across the brain, making it accessible to various cognitive processes.

In the GNWH context, electrophysiological markers such as mismatch negativity (MMN) or late positive component (P3b) are often considered potential NCC. For example, Dellert et al. showed that conscious perception of stimuli may be associated with the Visual Awareness Negativity (VAN) component but not always with enhanced P3b, questioning the universality of P3b as a direct consciousness correlate [Dellert et al., 2022]. This underscores the complexity of identifying unequivocal electrophysiological markers of consciousness and the need for further refinement.

However, despite all these advances, the NCC problem remains open. Daniel Dennett, for example, criticizes the idea of searching for a single “center” of consciousness or a unique neural correlate, arguing that consciousness is a distributed process arising from interactions of many parallel brain processes [Dennett]. He warns against the “Cartesian theater” notion, where a homunculus observes conscious experience. Patricia Churchland, in turn, emphasizes that consciousness is a brain property that will be fully explained by neurobiology without introducing any dualistic entities [Churchland].

The search for NCC is not merely about finding “hot spots” on brain scans but about understanding dynamic patterns, information integration, and global signal propagation underlying subjective experience. We see that various theories and methods offer unique perspectives, from quantitative integrated information indices to complexity analysis of neural signals in altered states. However, how can we be sure that these laboratory measurements truly reflect real conscious experience in everyday life beyond strictly controlled experiments?

Ecological Validity and Applicability of Consciousness Research

Research on neural correlates of consciousness (NCC), as discussed earlier, has undoubtedly advanced our understanding of how the brain generates subjective experience. However, when moving from laboratory conditions to the real world, the question arises: how applicable are these results beyond strictly controlled experiments? Here we face the problem of ecological validity, which questions the universality of conclusions drawn under artificially created conditions. Mudrik explicitly points this out, arguing that this step is critical for consciousness study, where experimental paradigms are typically artificial and small effect sizes are relatively common. In other words, to understand consciousness in its fullness, we need to go beyond simplified models.

The problem of artificiality in experimental paradigms is not new to cognitive sciences but becomes especially acute in consciousness research. Most NCC studies use stimuli far from natural: light flashes, simple images, brief sounds. These stimuli are carefully controlled to isolate specific neural responses but deprive conscious experience of its richness and context. Mudrik notes that consciousness research could greatly benefit from adopting a more ecological approach, similar to that already used in other cognitive science fields. Such an approach, in her view, can not only challenge existing hypotheses but also lead to discovering stronger effects and posing new research questions.

Take, for example, functional magnetic resonance imaging (fMRI), a cornerstone in NCC studies. Jianyang Wang et al. describe fMRI as a non-invasive neuroimaging method with high spatial resolution, allowing precise brain function localization. This is undoubtedly a powerful tool for detecting hidden consciousness not manifested through clinical behavior. However, as the same authors note, fMRI is susceptible to motion artifacts, has low temporal resolution, and is costly. Moreover, it is unsuitable for patients with disorders of consciousness (DoC) in intensive care units, limiting its applicability in real clinical scenarios.

Positron emission tomography (PET) is also used in DoC studies, measuring glucose metabolism, oxygen consumption, and neurotransmitter distribution [Wang et al., 2023]. These markers allow assessing residual brain function. However, as with fMRI, most PET-based studies, especially task-based paradigms, face the problem of small sample sizes and single cases, complicating result generalization [Wang et al., 2023]. This highlights the need for more extensive clinically relevant neuroimaging research.

When discussing disorders of consciousness such as vegetative state (VS/UWS) or minimally conscious state (MCS), ecological validity becomes even more critical. Studies show that VS/UWS and MCS patients may retain brain responses to linguistic and auditory stimuli detectable by fMRI. For example, Okumura et al. found bilateral temporal lobe activation in VS/UWS patients after musical stimulation, correlating with consciousness recovery. Wang et al. demonstrated that types and volumes of auditory cortex activation induced by familiar voices significantly correlated with prognosis in VS/UWS patients. However, although important, these studies are still conducted under controlled conditions, and how these responses manifest in everyday, non-laboratory environments remains an open question.

Moreover, even within fMRI studies, different approaches exist. Besides task-based paradigms, resting-state fMRI is actively used, which does not require interaction with DoC patients or complex experimental setups. Luppi et al. found that human consciousness depends on spatiotemporal interactions between brain integration and functional diversity by comparing resting-state fMRI data from awake volunteers, volunteers under propofol, and DoC patients. Huang et al. and Qin et al. also identified key brain networks, such as the default mode network (DMN) and dorsal attention network, playing important roles in maintaining consciousness. However, despite advantages, resting-state fMRI remains a laboratory method, and its results require cautious interpretation in real-life contexts.

While neuroimaging provides valuable data on brain activity, it cannot always capture subtle nuances of subjective experience. Daniel Dennett, for example, criticizes attempts to reduce consciousness to simple neural correlates, arguing that such an approach overlooks the complexity and multifaceted nature of the phenomenon. He would likely point out that even if we can measure brain activity associated with awareness, it does not mean we fully understand the experience itself.

On the other hand, David Chalmers, with his “hard problem of consciousness,” emphasizes that even perfect understanding of neural correlates will not explain why subjective experience exists at all. This philosophical distinction between “easy” and “hard” problems of consciousness directly influences how we assess the applicability of our research. If we focus only on the “easy” problems — mechanisms underlying consciousness — ecological validity may be less critical. But if we aim to understand the phenomenon of subjective experience itself, artificial experimental conditions become a serious obstacle.

In the context of artificial intelligence (AI) and potential emergence of artificial consciousness, the problem of ecological validity transforms into the search for “ecumenical heuristics.” Shevlin proposes seeking heuristics that allow preliminary assessments of the likelihood of consciousness emergence in various artificial systems. This means developing criteria independent of specific biological implementation but capable of capturing essential aspects of conscious experience. Essentially, this is an attempt to create a universal “ecological” test for consciousness applicable to both biological and artificial systems.

Although neuroimaging methods such as fMRI and PET provide unprecedented opportunities for studying neural correlates of consciousness, their ecological validity and applicability in real-world conditions remain subjects of active debate. The need to move toward more naturalistic and contextually rich experimental paradigms, as well as to develop universal criteria for consciousness assessment, are key tasks for the future. Without this, we risk gaining deep understanding of consciousness in vitro but remaining ignorant of its manifestations in everyday life and its potential emergence in non-biological systems. This leads us to critical reflection on the limitations of current approaches and the need for their revision.

Criticism and Limitations

Methodological Limitations and the Problem of Ecological Validity

Despite significant progress in consciousness research, current methods and experiments face fundamental limitations that question the universality and applicability of obtained results. One key problem is the low ecological validity of many experimental paradigms. As Mudrik et al. note, consciousness studies often use artificial stimuli and laboratory conditions far removed from everyday experience. This leads to experimental paradigms increasingly diverging from everyday conscious and unconscious processes, raising concerns about their applicability to real life [Mudrik et al., 2024]. For example, studying neural correlates of consciousness (NCC) using binocular rivalry or masking, while allowing sensory input control, does not reflect the complexity and dynamism of natural perception. What if consciousness in its fullness manifests only in rich, multidimensional interaction with the world, not in response to isolated stimuli?

Another limitation relates to the very nature of measuring consciousness. Most methods, whether neuroimaging or behavioral tests, essentially measure access consciousness — the ability of information to be used for reasoning, speech, or action — rather than phenomenal consciousness — subjective experience, what it is like to be in that state [Block, 1995]. Even if we detect brain activity correlating with conscious perception, this does not guarantee subjective experience. For example, blindsight patients can correctly indicate object locations without awareness, demonstrating information processing without phenomenal consciousness [Block, 1995]. This creates a methodological trap: we can measure only what manifests but not what is experienced. How then can we be sure that our “objective” measures of consciousness truly reflect subjective experience rather than complex cognitive processes that may occur without it?

The Problem of Interpreting Neural Correlates of Consciousness

Even if we successfully identify neural correlates of consciousness (NCC), their interpretation remains a subject of intense debate. Are these correlates causes of consciousness, its consequences, or merely accompanying phenomena? As Sandberg et al. emphasize, traditional contrastive analyses comparing brain activity during conscious and unconscious perception may conflate true NCC with precursors or consequences of consciousness. For example, prefrontal cortex activation may relate not to conscious experience itself but to decision-making or reporting processes [Dellert et al., 2022]. If we could isolate true NCC, we might find they are either more distributed or more localized than current models suggest.

Moreover, there is the “hard problem of consciousness,” formulated by David Chalmers. No amount of neural activity data, even if perfectly correlated with conscious experience, can explain why this activity produces subjective experience. Neuroimaging methods like fMRI and EEG can show where and when neural events related to consciousness occur but do not answer how physical processes become qualia. For example, Jianyang Wang et al. note that fMRI has high spatial but low temporal resolution, while EEG is the opposite. Multimodal approaches attempt to compensate for these shortcomings but cannot bridge the fundamental gap between objective description and subjective experience. This means that despite all technological advances, we may end up with a complete map of neural correlates yet still not know why they generate consciousness at all.

Conclusions

• Consciousness is a multidimensional phenomenon requiring an interdisciplinary approach integrating neuroscience, psychology, philosophy, and informatics.
• The distinction between “easy” and “hard” problems of consciousness remains central to research, focusing on subjective experience (qualia) as a key unresolved issue.
• Neural correlates of consciousness (NCC) are identified by comparing brain activity during conscious and unconscious stimulus perception, but their precise nature and localization remain debated.
• Integrated Information Theory (IIT) and Global Neuronal Workspace Theory (GNWT) offer competing yet complementary explanations of consciousness architecture and functions, each with strengths and limitations.
• Modern neuroimaging methods such as fMRI and EEG are key to analyzing neural processes of consciousness, but their ecological validity and real-world applicability remain questionable.
• Studies of altered and disordered states of consciousness provide valuable data on the flexibility and dynamics of conscious experience, but their interpretation requires caution and philosophical consideration.
• How can we develop universal, biologically independent criteria for consciousness assessment applicable to both biological and artificial systems, and what would this mean for our understanding of the phenomenon itself?