Although very technical at times, this article is very valuable in describing the different mood disorders and other conditions that lead to Self-injury.
In the DSM-IV, the only diagnoses that mention self-injury as a symptom or criterion for diagnosis are borderline personality disorder, stereotypic movement disorder (associated with autism and mental retardation), and factitious (faked) disorders in which an attempt to fake physical illness is present (APA, 1995; Fauman, 1994). It also seems to be generally accepted that extreme forms of self-mutilation (amputations, castrations, etc) are possible in psychotic or delusional patients. Reading the DSM, one can easily get the impression that people who self-injure are doing it willfully, in order to fake illness or be dramatic. Another indication of how the therapeutic community views those who harm themselves is seen in the opening sentence of Malon and Berardi's 1987 paper "Hypnosis and Self-Cutters":

Since self-cutters were first reported on in 1960, they have continued to be a prevalent mental health problem. (emphasis added)

To these researchers, self-cutting is not the problem, the self-cutters are.

However, self-injurious behavior is seen in patients with many more diagnoses than the DSM suggests. In interviews, people who engage in repetitive self-injury have reported being diagnosed with depression, bipolar disorder, anorexia, bulimia, obsessive-compulsive disorder, post-traumatic stress disorder, many of the dissociative disorders (including depersonalization disorder, dissociative disorder not otherwise specified, and MPD/DID), anxiety and panic disorders, and impulse-control disorder not otherwise specified. In addition, the call for a separate diagnosis for self-injurers is being taken up by many practitioners.

It is beyond the scope of this page to provide definitive information about all of these conditions. I will try, instead, to give a basic description of the disorder, explain when I can how self-injury might fit into the pattern of the disease, and give references to pages where much more information is available. In the case of borderline personality disorder (BPD), I devote considerable space to discussion simply because the label BPD is sometimes automatically applied in cases where self-injury is present, and the negative effects of a BPD misdiagnosis can be extreme.


As mentioned, self-injury is often seen in those with autism or mental retardation; you can find a good discussion of self-harm behaviors in this group of disorders at The Center for the Study of Autism.


"Every time I say something they find hard to hear, they chalk it up to my anger, and never to their own fear."


Unfortunately, the most popular diagnosis assigned to anyone who self-injures is borderline personality disorder. Patients with this diagnosis are frequently treated as outcasts by psychiatrists; Herman (1992) tells of a psychiatric resident who asked his supervising therapist how to treat borderlines was told, "You refer them." Miller (1994) notes that those diagnosed as borderline are often seen as being responsible for their own pain, more so than patients in any other diagnostic category. BPD diagnoses are sometimes used as a way to "flag" certain patients, to indicate to future care givers that someone is difficult or a troublemaker. I sometimes used to think of BPD as standing for "Bitch Pissed Doc."
This is not to say that BPD is a fictional illness; I have encountered people who meet the DSM criteria for BPD. They tend to be people in great pain who are struggling to survive however they can, and they often unintentionally cause great pain for those who love them. But I have met many more people who don't meet the criteria but have been given the label because of their self-injury.
Consider, however, the DSM-IV Handbook of Differential Diagnosis (First et al. 1995). In its decision tree for the symptom "self-mutilation," the first decision point is "Motivation is to decrease dysphoria, vent angry feelings, or to reduce feelings of numbness... in association with a pattern of impulsivity and identity disturbance." If this is true, then a practitioner following this manual would have to diagnose someone as BPD purely because they cope with overwhelming feelings by self-injuring.
This is particularly disturbing in light of recent findings (Herpertz, et al., 1997) that only 48% of their sample of self-injurers met the DSM criteria for BPD. When self-injury was excluded as a factor, only 28% of the sample met the criteria.
Similar results were seen in a 1992 study by Rusch, Guastello, and Mason. They examined 89 psychiatric inpatients who had been diagnosed as BPD, and summarized their results statistically.
Different raters examined the patients and the hospital records and indicated the degree to which each of the eight defining BPD symptoms were present. One fascinating note: only 36 of the 89 patients actually met the DSM-IIIR criteria (five of eight symptoms present) for being diagnosed with the disorder. Rusch and colleagues ran a statistical procedure called factor analysis in an effort to discover which symptoms tend to co-occur.
The results are interesting. They found three symptom complexes: the "volatility" factor, which consisted of inappropriate anger, unstable relationships, and impulsive behavior; the "self-destructive/unpredictable" factor, which consisted of self-harm and emotional instability; and the "identity disturbance" factor.
The SDU (self-destructive) factor was present in 82 of the patients, while the volatility was seen in only 25 and the identity disturbance in 21. The authors suggest that either self-mutilation is at the core of BPD or clinicians tend to use self-harm as a sufficient criterion to label a patient BPD. The latter seems more likely, given that fewer than half of the patients studied met the DSM criteria for BPD.

One of the foremost researchers into Borderline Personality Disorder, Marsha Linehan, does believe that it is a valid diagnosis, but in a 1995 article notes: "No diagnosis should be made unless the DSM-IV criteria are strictly applied. . . . the diagnosis of a personality disorder requires the understanding of a person's long-term pattern of functioning." (Linehan, et al. 1995, emphasis added.) That this does not happen is evident in the increasing numbers of teenagers being diagnosed as borderline. Given that the DSM-IV refers to personality disorders as longstanding patterns of behavior usually beginning in early adulthood, one wonders what justification is used for giving a 14-year-old a negative psychiatric label that will stay with her all of her life? Reading Linehan's work has caused some therapists to wonder if perhaps the label "BPD" is too stigmatized and too over-used, and if it might be better to call it what it really is: a disorder of emotional regulation.

If a care giver diagnoses you as BPD and you're fairly certain the label is inaccurate and counterproductive, find another doctor. Wakefield and Underwager (1994) point out that mental health professionals are no less likely to err and no less prone to the cognitive shortcuts we all take than anyone else is:

When many psychotherapists reach a conclusion about a person, not only do they ignore anything that questions or contradicts their conclusions, they actively fabricate and conjure up false statements or erroneous observations to support their conclusion [note that this process can be unconscious] (Arkes and Harkness 1980). When given information by a patient, therapists attend only to that which supports the conclusion they have already reached (Strohmer et al. 1990). . . . The frightening fact about conclusions reached by therapists with respect to patients is that they are made within 30 seconds to two or three minutes of the first contact (Ganton and Dickinson 1969; Meehl 1959; Weber et al. 1993). Once the conclusion is reached, mental health professionals are often impervious to any new information and persist in the label assigned very early in the process on the basis of minimal information, usually an idiosyncratic single cue (Rosenhan 1973) (emphasis added).

[NOTE: My inclusion of a quote from these authors does not constitute a full endorsement of their entire body of work.]

Mood Disorders

Self-injury is seen in patients who suffer from major depressive illness and from bipolar disorder. It is not exactly clear why this is so, although all three problems have been linked to deficiencies in the amount of serotonin available to the brain. It is important to separate the self-injury from the mood disorder; people who self-injure frequently come to learn that it is a quick and easy way of defusing great physical or psychological tension, and it is possible for the behavior to continue after the depression is resolved. Care should be taken to teach patients alternative ways to cope with distressing feelings and over-stimulation.
Both major depression and bipolar disorder are enormously complex diseases; for a thorough education on depression, go to

To find out more about bipolar disorder, try The Pendulum Resource Page, presented by members of one of the first mailing lists created for bipolar people.


Self-inflicted violence is often seen in women and girls with anorexia (a disease in which a person has an obsession with losing weight, dieting, or fasting, and as a distorted body image -- seeing his/her skeletal body as "fat") or bulimia (an eating disorder marked by binges where large amounts of food are eaten followed by purges, during which the person attempts to remove the food from her/his body by forced vomiting, abuse of laxatives, excessive exercise, etc).
There are many theories as to why SI and eating disorders co-occur so frequently. Cross is quoted in Favazza (1996) as saying that the two sorts of behavior are

attempts to own the body, to perceive it as self (not other), known (not uncharted and unpredictable), and impenetrable (not invaded or controlled from the outside. . . . [T]he metaphorical destruction between body and self collapses [ie, is no longer metaphorical]: thinness is self-sufficiency, bleeding emotional catharsis, bingeing is the assuaging of loneliness, and purging is the moral purification of self. (p.51)

Favazza himself favors the theory that young children identify with food, and thus during the early stages of life, eating could be seen as a consuming of something that is self and thus make the idea of self-mutilation easier to accept. He also notes that children can anger their parents by refusing to eat; this could be a prototype of self-mutilation done to retaliate against abusive adults. In addition, children can please their parents by eating what they are given, and in this Favazza sees the prototype for SI as manipulation.
He does note, though, that self-injury brings about a rapid release from tension, anxiety, racing thoughts, etc. This could be a motivation for an eating-disordered person to hurt him/herself -- shame or frustration at the eating behavior leads to increased tension and arousal and the person cuts or burns or hits to obtain quick relief from these uncomfortable feelings. Also, from having spoken to several people who both have an eating disorder and self-injure, I think it's quite possible that self-injury offers some an alternative to the disordered eating. Instead of fasting or purging, they cut.
There haven't been many laboratory studies probing the link between SI and eating disorders, so all of the above is speculation and conjecture.
Two eating-disorders web pages -- the ED section of Something Fishy and its associated site, Mirror, Mirror -- are probably the best sources for detailed information on eating disorders.


Self-injury among those diagnosed with OCD is considered by many to be limited to compulsive hair-pulling (known as trichotillomania and usually involving eyebrows, eyelashes, and other body hair in addition to head hair) and/or compulsive skin picking/scratching/excoriation. In the DSM-IV, though, trichotillomania is classified as an impulse-control disorder, and OCD as an anxiety disorder. Unless the self-injury is part of a compulsive ritual designed to ward off some bad thing that would otherwise happen, it should not be considered a symptom of OCD. The DSM-IV diagnosis of OCD requires:

  1. the presence of obsessions (recurrent and persistent thoughts that are not simply worries about everyday matters) and/or compulsions (repetitive behaviors that a person feels a need to perform (counting, checking, washing, ordering, etc) in order to stave off anxiety or disaster);
  2. recognition at some point that the obsessions or compulsions are unreasonable;
  3. excessive time spent on obsessions or compulsions, reduction of quality of life due to them, or marked distress due to them;
  4. the content of the behaviors/thoughts is not confined to that associated with any other Axis I disorder currently present;
  5. the behavior/thoughts not being a direct result of medication or other drug use.

The current consensus seems to be that OCD is due to a serotonin imbalance in the brain; SSRI's are the drug of choice for this condition. A 1995 study of self-injury among female OCD patients (Yaryura-Tobias et al.) showed that clomipramine (a tricyclic antidepressant known as Anafranil) reduced the frequency of both compulsive behaviors and of SIB. It is possible that this reduction came about simply because the self-injury was a compulsive behavior with different roots than SIB in non-OCD patients, but the study subjects had much in common with them -- 70 percent of them had been sexually abused as children, they showed the presence of eating disorders, etc. The study strongly suggests, again, that self-injury and the serotonergic system are somehow related.
Post-Traumatic Stress Disorder

Post-Traumatic Stress Disorder refers to a collection of symptoms that may occur as a delayed response to a serious trauma (or series of traumas). More information on the concept is available in my quick Trauma/PTSD FAQ. It's not meant to be comprehensive, but just to give an idea of what trauma is and what PTSD is about. Herman (1992) suggests an expansion of the PTSD diagnosis for those who have been continually traumatized over a period of months or years. Based on patterns of history and symptomology in her clients, she created the concept of Complex Post-Traumatic Stress Disorder. CPTSD includes self-injury as a symptom of the disordered affect regulation severely traumatized patients often have (interestingly enough, one of the main reasons people who hurt themselves do so is in order to control seemingly uncontrollable and frightening emotions). This diagnosis, unlike BPD, centers on why patients who self-harm do so, referring to definite traumatic events in the client's past. Although CPTSD is not a one-size-fits-all diagnosis for self-injury any more than BPD is, Herman's book does help those who have a history of repeated severe trauma understand why they have so much trouble regulating and expressing emotion.

Cauwels (1992) calls PTSD "BPD's identical cousin." Herman seems to favor a view in which PTSD has been fragmented into three separate diagnoses:

Area of most prominent dysfunction Diagnosis given
Somatic/physioneurotic (Bodily dysregulation -- problems regulating or understanding messages from the body and/or expression of emotional distress in physical symptoms) Conversion Disorder
(formerly Hysterical Neurosis)
Consciousness Deformation (breakdown in the ability to perceive oneself as a single entity with an uninterrupted history or to integrate body and consciousness) Dissociative Identity Disorder/
Multiple Personality Disorder
Dysregulation of identity, emotions, and relationships Borderline Personality Disorder

For an incredible amount of information on trauma and its effects, including post-trauma stress syndromes, definitely visit David Baldwin's Trauma Information Pages.


The dissociative disorders involve problems of consciousness -- amnesia, fragmented consciousness (as seen in DID), and deformation or alteration of consciousness (as in Depersonalization Disorder or Dissociative Disorder Not Otherwise Specified ).
Dissociation refers to a sort of turning off of consciousness. Even psychologically normal people do it all the time -- the classic example is a person who drives to a destination while "zoning out" and arrives not remembering much at all about the drive. Fauman (1994) defines it as "the splitting off of a group of mental processes from conscious awareness." In the dissociative disorders, this splitting off has become extreme and often beyond the patient's control.


Depersonalization is a variety of dissociation in which one suddenly feels detached from one's own body, sometimes as if they were observing events from outside themselves. It can be a frightening feeling, and it may be accompanied by a lessening of sensory input -- sounds may be muffled, things may look strange, etc. It feels as if the body is not part of the self, although reality testing remains intact. Some describe depersonalization as feeling dreamlike or mechanical. A diagnosis of depersonalization disorder is made when a client suffers from frequent and severe episodes of depersonalization. Some people react to depersonalization episodes by inflicting physical harm on themselves in an attempt to stop the unreal feelings, hoping that pain will bring them back to awareness. This is a common reason for SI in people who dissociate frequently in other ways.


DDNOS is a diagnosis given to people who show some of the symptoms of other dissociative disorders but do not meet the diagnostic criteria for any of them. A person who felt she had alternate personalities but in whom those personalities were not fully developed or autonomous or who was always the personality in control might be diagnosed DDNOS, as might someone who suffered depersonalization episodes but not of the length and severity required for diagnosis. It can also be a diagnosis given to someone who dissociates frequently without feeling unreal or having alternate personalities. It's basically a way of saying "You have a problem with dissociation that affects your life negatively, but we don't have a name for exactly the sort of dissociation you do." Again, people who have DDNOS often self-injure in an attempt to cause themselves pain and thus end the dissociative episode.


In DID, a person has at least two personalities who alternate taking full conscious control of the patients behavior, speech, etc. The DSM specifies that the two (or more) personalities must have distinctly different and relatively enduring ways of perceiving, thinking about, and relating to the outside world and to the self, and that at least two of these personalities must alternate control of the patient's actions.
DID is somewhat controversial, and some people claim that it is over-diagnosed. Therapists must be extremely careful in diagnosing DID, probing without suggesting and taking care not to mistake undeveloped personality facets for fully-developed separate personalities. Also, some people who feel as if they have "bits" of them that sometimes take over but always while they're consciously aware and able to affect their own actions may run a risk of being misdiagnosed as DID if they also dissociate.
When someone has DID, they may self-injure for any of the reasons other people do. They may have an angry alter who attempts to punish the group by damaging the body or who chooses self-injury as a way of venting his/her anger.
It's extremely important that diagnoses of DID be made only by qualified professionals after lengthy interviews and examinations. For reliable information on all aspects of dissociation including DID, the International Society for the Study of Dissociation web site and The Sidran Foundation are good sources.

Kirsti's essay on "bits" provides reassuring and valuable information about DDNOS, the space between normal daydreaming and being DID.


The DSM groups many disorders under the heading of "Anxiety Disorders." The symptoms and diagnoses of these vary greatly, and sometimes people with them use self-injury as a self-soothing coping mechanism. They've found that it brings fast temporary relief from the incredible tension and arousal that build up as they grow progressively more anxious. For a good selection of writings and links about anxiety, try tAPir (the Anxiety-Panic internet resource).


I include this diagnosis simply because it is becoming a preferred diagnosis for self-injurers among some clinicians. This makes excellent sense when you consider that the defining criteria of any impulse-control disorder are (APA, 1995):

  • Failure to resist an impulse, drive, or temptation to perform some act that is harmful to the person or others. There may or may not be conscious resistance to the impulse. The act may or may not be planned.
  • An increasing sense of tension or [physiological or psychological] arousal before committing the act.
  • An experience of either pleasure, gratification, or release at the time of committing the act. The act . . . is consistent with the immediate conscious wish of the individual. Immediately following the act there may or may not be genuine regret, self-reproach, or guilt.

This describes the cycle of self-injury for many of the people I've talked to.


Favazza and Rosenthal, in a 1993 article in Hospital and Community Psychiatry, suggest defining self-injury as a disease and not merely a symptom. They created a diagnostic category called Repetitive Self-Harm Syndrome. This would be an Axis I impulse-control syndrome (similar to OCD), not an Axis II personality disorder. Favazza (1996) pursues this idea further in Bodies Under Siege. Given that it often occurs without any apparent disease and sometimes persists after other symptoms of a particular psychological disorder have subsided, it makes sense to finally recognize that self-injury can and does become a disorder in its own right. Alderman (1997) also advocates recognizing self-inflicted violence as a disease rather than a symptom.

Miller (1994) suggests that many self-harmers suffer from what she calls Trauma Reenactment Syndrome. Miller proposes that women who've been traumatized suffer a sort of internal split of consciousness; when they go into a self-harming episode, their conscious and subconscious minds take on three roles: the abuser (the one who harms), the victim, and the non-protecting bystander. Favazza, Alderman, Herman (1992) and Miller suggest that, contrary to popular therapeutic opinion, there is hope for those who self-injure. Whether self-injury occurs in concert with another disorder or alone, there are effective ways of treating those who harm themselves and helping them find more productive ways of coping.


Request a Refill

3 + 0 =
Solve this simple math problem and enter the result. E.g. for 1+3, enter 4.