Various conditions can lead to an alteration or even loss of consciousness. The origin can be physiological processes like sedatives, impairment of the brain by drugs (psychedelics, alcohol, sedatives, narcotics), or metabolic, ischemic or traumatic brain injury (Young, 2009; Eapen et al., 2017). With increasing severity this becomes a medical issue, particularly in the form of general anesthesia or coma. The extent can be scored according to residual responsiveness, e.g., by Glasgow Coma Scale, Ramsay-Scale, or Sedation-Agitation-Scale (Bordini et al., 2010). With total unconsciousness defined as the lack of any reaction to external stimuli, communication usually comes to an end, both from the patient’s and the helpers` side. But is this really so, and is it reasonable? Does unresponsiveness exclude perception? In case of doubt, communication should not be discontinued.

Several observations of emergency, surgical or intensive care patients suggest perception even in unconscious patients. When hypnosis is the establishment of contact with the unconscious and the influencing of psychological and physiological functions by means of suggestions (Erickson, 2009; Elkins et al., 2015; Linden et al., 2024), and if the possibility for this is opened up primarily by bypassing waking consciousness and the elimination of critical reason (Ahlskog, 2018; Peter, 2024, this issue), then the question arises as to whether hypnotic suggestions also have an effect when waking consciousness is otherwise eliminated, namely by coma or general anesthesia. Touching the unconscious in the unconscious.

Considering these situations and conditions it is important to remember that perception and its impact largely depend on the importance and meaning of the perceived signal or message. Moreover, suggestibility, i.e., the extend of the reactions to suggestions, is massively increased in trance, a non-ordinary state of consciousness that is induced by hypnosis or extreme situations (Peter, 2024, this issue). The events that lead to the states to be discussed, namely an accident or a trauma resulting in coma, the need for surgery under general anesthesia, or complications and disorders making intensive care necessary, all represent such “extreme,” trance-inducing conditions. Could the triggered elevation in suggestibility also be significant when in the course of such events unconsciousness has occurred?

Before evaluating and discussing evidence for perception in different states and disorders of consciousness (DoC) and the possibility of communication, some definitions or rather what is understood by this in the following seem appropriate. “Consciousness” is a subjective experience that plays a considerable role in the psychological, physical, and behavioral reactions to external and internal stimuli. However, the precise definition can vary considerably between philosophers, neurophysiologists and clinicians. Here, clinical aspects are of priority. “Connectiveness” refers to the connection of consciousness to the external world allowing experience of external stimuli (Sanders et al., 2012). Examples of disconnected consciousness are dreaming in sleep, namely rapid-eye-movement (REM) sleep, or dreaming in anesthesia. It also can be induced by hypnosis. “Memory” is not essential for experience, nor is it for consequences of perception. For recalls of events that have taken place during unconsciousness a distinction is made between explicit and implicit memory. Explicit memories are reported spontaneously by the awakened patient, or can be determined by structured interviews or questionnaires after the phase of unconsciousness. Implicit memories are not consciously remembered by the patient, but can be evaluated under hypnosis (Levinson, 1965; Cheek, 1966), or by association techniques (Schwender et al., 1994). Remembrance, however, is strongly dependent on attention, emotional content and meaning. “Responsiveness” means the behavioral interaction with the outside world, and is divided in spontaneous and goal-directed (following a command) responses. This responsiveness is not only dependent on the perception of an input, but also limited by restrictions in the output, for instance by pharmacological muscle relaxation, psychic or neurologic paralysis, attention, and motivation. Another indicator for perception in unconscious patients is the phenomenon of near-death-experiences (NDE). They can be described as internal awareness experienced in unresponsive conditions and classified as disconnected consciousness (Martial et al., 2020). Such memories are reported after situations close to death, e.g., cardiac arrest or coma, characteristically connected with unconsciousness. The reported memories can be detailed and true, or false (Martial et al., 2018). Both harbor the risk of traumatization, e.g., the true perception of fixation straps during intensive care or the oneiroid “false memory” of being a war prisoner.

Clinically, mainly behavioral responses and memories are used for assessment of consciousness, which both neither allow precise judgment of consciousness, nor evaluation of the consequences of any perception. Even fragments of information can be behaviorally functional yet kept out of consciousness (Mashour, 2013).

Considering the effects of external stimuli including communication on patients that present as “unconscious,” one has to deal with all these components in its different appearance and characteristics, and their combinations. This creates great complexity and hampers simple equations like unresponsiveness = unconsciousness (Sanders et al., 2012). Moreover, this complexity is the reason for many disadvantageous misconceptions in a number of severe medical situations (Table 1).

The “cognitive unbinding paradigm” is based on the “integrated information” theory that describes unconsciousness as interrupted information. Consciousness is lost due to impaired communication across brain networks and the consequent isolation of cognitive processing modules (Mashour, 2013). Thus, isolation rather than extinction of neural activity or sensory processing is sufficient for unconsciousness. Or other way around, areas and networks in the brain involved in information synthesis and inter-modal processing may be disrupted, while sensory networks and processing can persist despite unconsciousness. Moreover, from hypnosis we know that psychological and physiological functions are especially regulated and modifiable in the unconscious mind, and reaching the conscious level is not necessary for effects of verbal and nonverbal suggestions (Knafo and Weinberger, 2024; Peter, 2024, this issue). Specific rather than general neural network disruption is the common cause of unconsciousness, with differing resilience of brain areas and their connections affected by trauma, circulatory disorder or drugs. Neuroimaging has demonstrated persistent sensory processing during impaired network communication in coma or general anesthesia. Functional connectivity of sensory networks was found relatively unperturbed for instance after anesthetics (Mashour, 2013). Cognitive processing can persist in unconscious states, while the binding of this activity into a meaningful conscious representation is inhibited, which on the other hand is not required to trigger effects. A summary of connections between consciousness and perception are depicted in Figure 1.

One of the most exciting reports of positive communication in emergency medicine is the “Kansas experiment.” There, avoidance of unrelated or negative conversation and a positive text recited during transport to the hospital resulted in more patients surviving the transport and the hospital stay, and quicker recovery rates (Jacobs, 1991). The hypnotherapist M.E. Wright had trained three groups of ambulance attendants to do so for 6 month and compared outcome to control groups. In this study, the text (Table 2) was used for accident victims regardless “whether they were stuporous, conscious, or unconscious,” which means that unconscious patients were included. It was assumed that with the trauma “a narrowing of the total psychological functioning has occurred so that there is an acute responsiveness in some areas and a lack of awareness in others,” and that “shock can be considered a radical mobilization of the body to preserve essential life functions to sustain survival” maintaining minimal reception of information. Wright had the idea that “in such situations the person’s usual critical responsiveness to the environment has been altered so that whatever stimuli do reach are often subject to literal translation and can either aggravate or support the life systems that are hanging on...” This description also perfectly applies to other forms of unconsciousness (see the following) and to states of “natural trance” induced by stress, fear and pain in emergencies or when facing surgery (Cheek, 1962a). It is a pity and a shame that in the time after the “Kansas experiment” of 1976, this study was never reproduced in the subsequent 48 years to be published in a medical journal.

The most important and convincing evidence for perception in unconscious emergency patients stems from studies on cardiopulmonary resuscitations (CPR). With his book “Life after life” of 1975 based on 150 interviews, Raymond Moody shaped the expression “near-death experience” and raised the public and scientific interest in this topic (Moody, 2001). In 1979 a study was published on 2000 patients interviewed after life-threatening situations. 60% reported near-death-experiences (NDE) including perception of the external processes and the stress of not being able to make themselves heard (Shoonmaker, 1979). In addition to the selective evaluation of patients in appropriate situations, there exist also epidemiological studies using representative surveys. NDE were found in 15% of US Americans, with one third reporting extraordinary experiences including out-of-body-experience (OBE), which usually is combined with perception and description of the external events. A survey of 4,000 Germans revealed NDE in 4.5% in the normal population, mainly after experiencing emergencies, surgeries or cardiac infarctions, with 6% actually suffering clinical death (Knoblauch et al., 2001). Particularly noteworthy is that 65% of them felt mentally wide awake, and 30.5% described OBEs.

More precise with regard to perception during death and CPR is the report of van Lommel et al. (2001). In a prospective study of 344 patients surviving hospital resuscitation after cardiac arrest, 18% had NDE, with no correlation to oxygen deficiency, duration of cardiac arrest and massage, medication, or religious belief. In some of these patients a flat EEG was recorded. 25% of the patients had an OBE and reported details of the CPR. One patient told the nurse exactly where she had placed his dental prosthesis lost during resuscitation. In another study, 90 out of 93 reports by patients after OBE were accurate (Holden et al., 2009). In 2014 the results of the AWARE (AWAREness during REsuscitation) study, a prospective study on awareness during CPR, was published (Parnia et al., 2014). Of 140 patients surviving in-hospital cardiac arrest and resuscitation (16% overall survival rate), 9% had NDE, while 2% described awareness with explicit recall of actual events related to their resuscitation, including seeing and hearing the rescuers. One had a verifiable period of conscious awareness “during which time cerebral function was not expected...” In the second prospective AWARE study 11 of 28 (21%) surviving and surveyable patients after resuscitation for cardiac arrest reported memories and perception from cardiac arrest without external signs of consciousness (Parnia et al., 2023). One described “they were putting two electrodes to my chest, and I remember the shock.”

All of these patients had suffered clinical death defined as the period of unconsciousness caused by total lack of oxygen in the brain (anoxia) resulting from the arrest of circulation, breathing, or both. Under the subsequent conditions of cardiac low flow during manual cardiac compressions the brain can survive but not function (van Lommel, 2011). It has been argued that effective CPR could allow temporary awareness. This is discrepant to the fact that even during effective CPR cardiac output and oxygenation are impaired and limited, which is incompatible with higher cerebral performances. The mentioned reports contrast completely to CPR-induced consciousness with observable signs, observed in 0.2–0.9% of resuscitations (West et al., 2022). On the other hand, an important characteristic of NDE is the high alertness and awareness reported by the patients, as well as the exceptionally good recall of the perceived experiences, even decades later. This reflects non-congruence of consciousness and perception. Furthermore, reports of synchronized gamma oscillations - signifying heightened lucid consciousness - in humans and animals on electroencephalography (EEG) during cardiac arrest and death, has raised the intriguing possibility of electrocortical biomarkers of lucid/heightened consciousness during cardiac arrest (Borjigin et al., 2013).

Another clue to perception could be the occurrence of posttraumatic stress disorder (PTSD) or other newly developed psychiatric morbidity after CPR such as depression, anxiety disorder, or substance abuse (Oh et al., 2022). How can somebody be stressed without being aware of the stressful situation? A high incidence of PTSD following CPR of 27% was reported (Gamper et al., 2004). Interestingly, sedation during CRP does not prevent PTSD but worsens survival.

Consciousness quite often is impaired in patients during intensive care due to coma after traumatic or ischemic brain injury, intoxication, or infectious or metabolic disorders, or due to pharmacological coma ranging from sedation to “medically-induced coma” (to reduce brain metabolism). The resulting disorders of consciousness (DoC) have a diverse appearance including coma, all sharing unresponsiveness (Hannawi et al., 2015). Coma is defined as a state of profound unawareness from which the closed-eyes, non-communicating patients cannot be aroused (Kondziella et al., 2020). In contrast, the term “vegetative state/unresponsive wakeful syndrome” denotes a condition of wakefulness without awareness, where patients open-eyes exhibit only reflex behaviors. These patients may recover to a “minimal conscious state” (MCS), where non-reflex cortically mediated behaviors fluctuate spontaneously or dependent on certain stimuli or specific situations. In addition, assessment of consciousness perception may be further obscured by existence of islands of consciousness and functional fluctuations. The situation is aggravated by an apparently innate reflex to stop communication when someone has their eyes closed.

The prevalence of PTSD after intensive care including such patients is high, amounting to 20-25% (Parker et al., 2015). The strains are manyfold: being fixed and restrained in bed, treated with vasopressors (that mimic stress response) or paralyzed, continuous noise and lighting, unpleasant manipulations, often mechanical ventilation, and the perceived severity of the life-threatening illness itself (Warlan and Howland, 2015). These strains are not restricted to conscious perception. For instance, PTSD signs and symptoms are found in 35% of mechanically ventilated patients that usually are comatose or sedated (Bienvenu et al., 2013). Rather than the clear memories of an awake patient, these are the delusional memories of frightening perceptual experiences that are associated with the development of PTSD, and are more likely to be retained over time (Jones et al., 2007). However, having no memory of ICU is not beneficial either (Granja et al., 2008). No memory of their admission to the ICU in half of the patients was found strongly associated with the development of PTSD. A considerable portion of these patients is sedated or unconscious at the time of admission, an important cause for their amnesia for that time. Moreover, the use of restrains, necessitated by unconscious agitation and movements, is associated with PTSD (Davydow et al., 2008).

The idea that psychological trauma is dependent on conscious perception and that sedation would reduce risk for PTSD has turned out to be wrong. In fact, the risk is enhanced by sedation. A review found use of benzodiazepines and duration of sedation correlating with PTSD (Wade et al., 2013). Possible explanations are that conscious perception reduces the trauma, or that the trauma is aggravated as communication is often stopped as soon as the patient’s eyes are closed, and he or she is ignored while inappropriate conversations may occur. The realization that drug-induced loss of consciousness is not protective, and can even enhance stress and PTSD led to attempts to reduce the stress of mechanical ventilation by lighter levels of sedation, intermittent spontaneous breathing trials and early extubation. However, their validation is yet missing. Instead, hypnotic communication has been used successfully to reduce stress and fasten weaning from the respirator (Szilágyi et al., 2014). Most suggestions, both negative and positive ones, especially those of company and care, are transmitted to and perceived by the unconscious mind. An example is the calming down of heart rate when a visiting relative speaks to a deeply sedated patient. Patients may be able to unconsciously gauge a nurse’s or doctor’s intention and sense, if they are stressed or compassionate. Difficult to frame into a study, most intensive care physicians can recall patients that after long recovery report events and words from times where they were considered unable to hear or perceive anything. Among other things, discussions and decisions are reported about to stop artificial coma (used to reduce brain metabolism during restricted blood flow with increased intracranial pressure) or to use extracorporal membrane oxygenation (ECMO, “artificial lung”). At the time of such decisions patients usually are deeply unconscious. Likewise, NDE have also been reported after severe brain injury of traumatic or other origin (Hou et al., 2013). Incidence was reported with 15%, and correlation to mechanical ventilation, sedation, surgical reason for admission, and dissociative propensity (Rousseau et al., 2023).

Severe acquired brain injuries resulting in a DoC provide a model from which insights into consciousness can be drawn (Di Perri et al., 2014). Diagnosis is difficult when based only on behavioral assessments, common in clinical routine. Latest research in terms of both improving the diagnosis of patients with DoC, and understanding the brain processes underlining consciousness, reveals an underlying broad and more complex than previously thought alteration of brain connectivity architecture. However, neuroimaging and electrophysical techniques are still insufficient to detect possible consciousness residuals in severely traumatic brain injured patients.

Interestingly, general anesthesia is also called drug-induced hypnosis, and literature searches for the term “hypnosis” regularly yield references about anesthesia. But can patients under anesthesia perceive anything at all, especially words? There are indications for it. David Cheek, an American gynecologist and hypnotherapist, was the first to point this out after narrations of his patients under hypnosis of events and conversations from earlier operations (Cheek, 1962b), without being believed. However, reports of intraoperative wakefulness with explicit memory increased, and to this day it occurs in about 0.2% of anesthetics (Ghoneim, 2000; Mashour and Avidan, 2015). Implicit memory can be revealed by association or under hypnosis. For example, in the 1990s Agnes Kaiser incorporated a text around the Robinson Crusoe story into a study using acoustic-evoked potentials (AEP) to investigate the influence of different anesthetic procedures on the primary auditory pathway as a way to measure depth of anesthesia (Schwender et al., 1994). The text played intraoperatively led postoperatively to associations Friday-holiday or Friday-island or Friday-Robinson instead of Friday-weekend-beginning. Effects of the positive text were unfortunately not studied at all. The observed high incidence is now considered to be due to inadequate anesthesia, and the actual incidence of implicit recall is reported to be 2%. Due to the negative content and consequences of such intraoperative perceptions, including a high incidence of PTSD, these are usually attributed to “inadequate anesthesia” and every effort is being made to avoid it. Modified anesthesia management and anesthesia depth monitoring have indeed reduced the traumatizing occurrence of intraoperative wakefulness, but has not yet been able to eliminate it (Tasbihgou et al., 2018). Moreover, primary sensory areas are relatively resistant to loss of consciousness under anesthesia (Nourski et al., 2018).

However, the recall of memories is shaped by meaning. What is important enough of intraoperative auditory stimuli to be remembered? A much higher incidence of intraoperative awareness was clearly demonstrated in the experiments of B.W. Levinson in South Africa in the 1960s, which today would be considered unethical (Levinson, 1965). In 10 patients under EEG-controlled general anesthesia a hypoxia alarm was simulated intraoperatively. “He′s got blue lips! There are ventilation difficulties...” and “I do not like this color!.” Postoperatively, four of the patients under hypnosis repeated the words correctly, while another four showed a fear reaction with termination of the trance. Accordingly, incidence of perception was 80%, in striking contrast to the otherwise reported occurrence of intraoperative awareness. Presumably, the reason for the high incidence in this case is the high, life-threatening significance of the given suggestion. Interestingly, massive efforts to rule out “intraoperative awareness” by EEG-derived monitors of anesthetic depth were not successful. The incidence can be reduced but not eliminated, its existence is not limited to insufficient anesthesia.

On the other hand, the phenomenon of intraoperative perception has repeatedly led to attempts to use it for positive suggestion. Some studies have reported reductions in pain, anxiety, postoperative nausea and vomiting (PONV), and subsequent need for medication (Williams et al., 1994; Nilsson et al., 2001). A recent meta-analysis identified 32 adequate randomized controlled trials (RCTs) from 7,427 reports involving 2,102 patients, but showed no effects on pain intensity or psychological distress, but small but significant positive effects on recovery and use of medication (Rosendahl et al., 2016). These findings raised hope that a non-pharmacologic approach such as therapeutic suggestion during general anesthesia might be beneficial for surgical patients. However, the RCTs identified were relatively old (1986-2001), small in size, and heterogeneous in design. In addition, therapeutic and prophylactic regimens have changed in the intervening period, the management and depth of anesthesia were not standardized in these studies, and the suggestions used were heterogeneous and often included negations. The approach also did not find its way into clinical routine anywhere.

However, recently the effect of hypnotic suggestions was investigated again. A controlled, randomized, triple-blinded multicenter study was conducted at five German university hospitals with 385 patients undergoing painful surgery of 1–3 h duration under general anesthesia (Nowak et al., 2020, 2022). A 20-min text set to background music, followed by a 10-min break, was played repeatedly over earphones for the entire duration of the surgery, and a text on anesthesia withdrawal was played in the final phase. The depth of anesthesia was strictly controlled and the intervention strictly during anesthesia only. The control group also received earphones but no audio recording. The intervention text was based on hypnotherapeutic principles, and did not contain negations (such as “they will not be in pain!”). Especially, issues of meaning such as competence and caring of the surgical and anesthesiologic team, self-regulation, dissociation to a safe place, affirmation, fear control, and confidence were addressed (Table 3) (Link to the text and audio file: https://www.frontiersin.org/articles/10.3389/fpsyg.2022.898326/full#supplementary-material). The results were a significant reduction in postoperative pain level (NRS) by 25% over the 24-h observation period. In line with this, a significantly reduced need for analgesics, namely the opioid piritramide and also the additional medication with non-opioids, by one third was observed. With 36.6% patients without any analgesics a number needed to treat (NNT) of six was found. This means that if six patients received this intraoperative communication, postoperative pain medication (including its potential side effects) was avoided entirely in one patient as a result of this treatment alone (Nowak et al., 2020). In these patients at increased risk for postoperative nausea and with vomiting (PONV), that common and debilitating side effect of surgery and anesthesia also was significantly reduced with the hypnotic intervention. Incidence of both early and late manifestations, i.e., early and delayed PONV, were halved. Moreover, an observed NNT of 7 indicates that medication with antiemetics can be avoided entirely in 1 of 7 patients (Nowak et al., 2022). This study demonstrates high efficacy in reducing side effects of anesthesia and surgery with a simple, practical, non-pharmacological intervention. In addition, it makes the case for a wide application of intraoperative hypnotic suggestions, as well as perioperative therapeutic communication for surgical patients in general.

EH: Writing – original draft, Writing – review & editing.