The following are some of the reasons for listing certain conditions, disorders, or syndromes as either “high-risk” or as “requiring caution” in the course of administering EEG biofeedback (neurofeedback).  These are not exhaustive and may be revised as the Board’s “High Risk Committee” continues deliberations on these matters.  The underlying assumption in all of these is that neurofeedback can be a powerful treatment modality in affecting symptoms associated with these conditions.  Accordingly, symptoms may be either decreased, as is most often the case, or they may be increased or exacerbated, particularly in conditions that may be characterized as manifesting paroxysmal behaviors secondary to instabilities in neuromodulatory networks.

High-risk conditions:

Seizures:

1. Perhaps the quintessential example of instability, seizures can be either ideopathic (unknown etiology) or caused by some insult to the CNS.  Moreover, they may be occur spontaneously (when it is impossible to determine the trigger), or they be triggered by causes outside the CNS (e.g., stressors, lights of a certain frequency).  The list of triggers is quite long and varies greatly among individuals and even within the same individual over time.  Even if a neurofeedback protocol does not directly cause a seizure, there is the possibility that it could trigger a cascade of events that could eventuate in a seizure.  Furthermore, something besides the neurofeedback protocol per se (e.g., something in the room) could trigger it, or it could occur spontaneously.  In any event, a practitioner must be familiar with these conditions, how to recognize seizures, what to do, and when to refer.  Untreated seizures often get worse over time, possibly due to such mechanisms as kindling.  The sequelae of seizures can vary greatly, ranging from minimal to death (often due to asphyxia).  However, even in cases where no apparent sequelae occur, it is likely that additional neurons are negatively affected, as seizure foci spread throughout the brain.  Moreover, hypoxic encephalopathy may occur if extended seizures occur.  It has been well established by neuropsychological assessments that many individuals with re-occurring seizures suffer cognitive deficits and often personality changes, particularly with temporal lobe epilepsy.

Bipolar Disorders:

1. Because these individuals vacillate between manic and depressive episodes, any treatment modality that can decrease one type of episode could potentially trigger the opposite episode.  Therefore, both manic and depressive episodes could be triggered by neurofeedback.  Mania could be manifested by psychosis, violence, or behavior that is dangerous to self (e.g., taking dangerous illicit drugs) or to one’s well being (spending sprees).  Depression could result in a suicidal attempt that could be fatal or that could result in severe brain damage.  Of course, even without such dire consequences, a depressive episode should be avoided.  It should also be mentioned that correct diagnosis can often be difficult, particularly the differential between substance dependence or abuse, borderline personality disorder, ADHD, and bipolar disorder.  Practitioners should be educated and trained in such matters before even attempting neurofeedback.

Major Depression:

1. Depressive symptoms could become worse with neurofeedback.
2. Suicidal attempts could occur if the depressive episode becomes sufficiently severe.  This must be understood even if the neurofeedback per se does not cause the worsening of the episode.
3. As the depressive episode decreases in severity, the probability of suicide could increase.  This can occur for several reasons, including but not limited to (a) an increase in energy, which can take an individual from a state of apathy and vegetative psychomotor retardation to a state of being able to engage in the extreme act of suicide; (b) a decrease in helplessness but an increase in hopelessness, so that one now feels able to do something about a hopeless life, namely end it; (c) a decrease in cognitive deficits (e.g., difficulty concentrating and making decisions) caused by the prior severe level of depression, so that one can now make plans (e.g., finalize one’s estate) and make the decision to suicide.

Current suicidal ideation or behavior or a history of suicidal attempts:

1. Of course, the above applies to any patient with suicidal ideation or behavior.  Note the inclusion of suicidal behavior.  That is, an individual with depression may not express such ideation, or the practitioner may not know how to make such assessments.  Then it is important that the clinician know some of the typical behaviors of individuals about to suicide.  A history of suicidal attempts is important, as the past is one of the best general predictors of the future, and it has been well established that a previous suicidal attempt is a predictor of a future attempt.  In addition, it has been established in the literature that depressive episodes tend to worsen over time.

Acquired brain injury (ABI):

1. Individuals with ABI, particularly traumatic brain injury (TBI), may be at risk for seizures; see above concerning seizures.
2. Such individuals are usually quite complex for numerous reasons, including but not limited to (a) their brains have suffered frank damage, which is often difficult to assess even with extensive neurological, neuroimaging, and neuropsychological techniques; (b) cognitive, personality, emotional, and behavioral sequelae usually occur; and (c) psychological, premorbid, and psychosocial factors often interact in complex ways.  If a practitioner does not know these variables (e.g., does not have information that could be obtained from a neuropsychological assessessment), he (she) is treating a complex condition with partial information.  At best, this would be considered substandard care; at worst, this could lead to protocols that could improve some symptoms but make others worse, especially if the clinician does not even know of these other areas.  As an example, a protocol being used for decreasing a headache in a TBI patient could trigger executive disinhibiton or acute anxiety, particularly if that patient has dysfunction in his frontal lobes. 
3. In view of the complexity cited under #2, the optimal treatment is very likely to be a combination of several modalities.  Since many neurofeedback clinicians appear to be treating ABI individuals with only neurofeedback and usually without much other knowledge from other professionals and often with little knowledge of this population, it is reasonable to list ABI as a high-risk condition. 

Conditions not considered “high-risk” but that justify extra caution:

Narcolepsy:

1. Individuals can be dangerous to themselves and others due to sudden losses of consciousness.  Because neurofeedback appears to affect sleep so frequently, it seems reasonable to assume that it could make this condition worse in a given case, even though history indicates that sleep usually normalizes with neurofeedback.  If neurofeedback has the potential of making such a condition worse, even if only temporarily, real harm could occur.    

Reactive Attachment Disorder (RAD):

1. There is the potential for triggering a period in individuals with RAD during which they could become violent and antisocial to the point that they could commit homicide or serious harm to property or others.  Serious decompensation could also occur with other symptoms becoming worse.

Dissociative Identity Disorder:

1. Due to the extreme complexity of this condition and how little is really understood about it, it should be considered a condition that warrants caution, if not “high-risk.”  This would probably pertain especially to alpha-theta protocols.  The primary risk appears to be the further splitting of the personality.  Another risk may be that an alpha-theta session could cause a severe abreaction with a clinician present who is not fully qualified to handle such states.     

Borderline Personality Disorder:

1. This condition is characterized by extensive instability of mood and behavior.  An individual with this condition may rapidly cycle between extreme states or poles, often much faster than someone with a bipolar disorder.  Such states often result in self-mutilating behavior and can be very harmful to the individual or to others.  The differential diagnosis with other conditions can be difficult, even with a well-trained clinician.  Again, perhaps this condition may even warrant being in the “high-risk” category, but we do not want too many conditions there. 

Panic Disorder:

1. There is the potential that a protocol could trigger a panic attack in someone so disposed.  Because these can occur so quickly, caution is in order. 

Migraine:

1. This appears to be the result of vessels suddenly dilating after having constricted.  Again, this is an example of instability.  If neurofeedback can stop a migraine attack during a session, there exists the potential of exacerbating one. 

Sleep apnea:

1. It may not be absolutely clear that sleep apnea is always a condition that "justifies extra caution", but certainly if sleep apnea gets worse, there is the real possibility of increased cognitive dysfunction, since it has been established that individuals with this condition do in fact suffer cognitive impairments.

Individuals with a history of psychotic episodes:

1. The risks will vary significantly from individual to individual.  Even if such an individual is clearly in remission and not currently psychotic, a neurofeedback session could trigger decompensation in someone who is at risk.