Clinical descriptions

Early descriptions of the borderline syndrome in children were replete with references to cognitive and developmental problems. Anna Freud (1969) proposed that children who would now be characterized as suffering from borderline disorder (BD) had massive developmental arrests, poor reality testing, and difficulty synthesizing information. Murray (1979) hypothesized that minimal brain dysfunction (MBD) was a major etiological factor in the development of BD, and that speech and language disorders were common. Similarly, Kernberg (1983) described the syndrome as at “the border” of an organic disorder and related to MBD. Bemporad, Smith, Hanson, and Cicchetti (1982) also identified “organic impairment” in these children, and described social awkwardness, neurological soft signs, and general unevenness in development. Pine (1983, 1986) was the first to hypothesize that constitutional neuropsychological deficits in combination with early trauma underlie the child's problems and interfere with normal development. He observed that children with this syndrome exhibited excessive dependence on the presence of a primary object to maintain stable functioning. These clinical observations led to a conceptualization of the syndrome as an atypical pervasive developmental disorder (PDD) with disturbances in cognitive processes, social relations, anxiety, attentional regulation, and neuromaturation (Cohen, Shaywitz, Young, & Shaywitz, 1983).

More recently, BD has been characterized by complex developmental deficits and integrative failures or a complex developmental (socioemotive–dyssocial) disorder (Ad-Dab'Bagh & Greenfield, 2001; Lincoln, Bloom, Katz, & Boksenbaum, 1998; Towbin, Dykens, Pearson, & Cohen, 1993). This conceptualization has evolved into use of the term multiple complex developmental disorder (MCDD) under the rubric of PDD with an onset before age 6 and characterized by functional deficits that lead to a consistent, enduring pattern of (a) impaired regulation of affective state, (b) consistent impairments in social behavior and sensitivity, and (c) impaired cognitive processing (Cohen, Paul, & Volkmar, 1986; Towbin et al., 1993).

A cluster analytic study of four groups of children found that fluctuations in the level of functioning was the strongest differentiating feature between children with MCDD and other childhood disorders (Van der Gaag et al., 1995). This feature, “fluctuations in level of functioning” corresponds to the DSM-IV description of “pervasive instability” supporting a similarity between the child and adult disorder.

Empirical studies of cognitive problems in children

The studies that have formally investigated children with the borderline syndrome have consistently found evidence for neurocognitive impairment. Hartocollis (1968) found that 15 adolescents diagnosed with the borderline syndrome exhibited moderate impairments on a modified Halstead–Reitan neuropsychological battery. A prospective study that followed children diagnosed with MBD or learning disorders into adulthood found that those diagnosed with the borderline syndrome exhibited language and thinking deficits (Milman, 1979). An in-depth review of 106 adolescents and adult patients found that 40% had a history of organicity (Andrulonis et al., 1980). Other studies have found that children with BD have a variety of heterogeneous cognitive impairments (Bemporad et al., 1982) and heterogeneous developmental learning disorders including auditory processing and discrimination problems (Berg, 1992; Gualtieri, Koriath, & Van Bourgondien 1983).

Two recent controlled studies provide more compelling evidence for cognitive problems in children with BD. One studied the pathophysiology of children in a residential treatment setting with BD/MCDD and comorbid attention-deficit/hyperactivity disorder (ADHD) compared to children with ADHD alone and normal controls (Lincoln et al., 1998). The borderline group was distinguished by the presence of wide fluctuations in functioning. They exhibited auditory processing abnormalities in both nonlinguistic and linguistic information processing and impairment on nonverbal measures of executive functioning involving planning and cognitive flexibility.

Lincoln et al.'s study (1998) also investigated attentional performance in all three groups using an event-related potentials (ERPs) paradigm. The BD group performance was unique when compared to ADHD and normal controls in that they exhibited consistently smaller amplitudes when presented with rare auditory but not visual targets, suggesting increased difficulty processing auditory versus visual stimuli. The latency of P300 was also longer in both the BD/MCDD and ADHD children compared to normal controls.

Another neuropsychological study compared BD children to those with other primary diagnoses in a day treatment program (Paris, Zelkowitz, Guzder, Joseph, & Feldman, 1999). Both groups exhibited neurological soft signs and impaired performance on visual memory and planning, vigilance, and attention, but the borderline group had a unique profile with problems in cognitive flexibility and abstraction as measured by the Wisconsin Card Sort Test (WCST; Heaton, 1981). The BD group appeared unable to achieve an overall conceptualization of the demands set by the test. These findings were independent of comorbidity with conduct disorder and ADHD.

A study supporting the theory of a transaction among neurocognitive and psychosocial factors in the development of the disorder compared the relative contribution of psychosocial risk factors and a set of neuropsychological variables (Zelkowitz, Paris, Guzder, & Feldman, 2001). The authors found that both psychological trauma and deficits in executive functioning act as risk factors for borderline pathology in children and make separate and independent contributions. The combination of both factors correctly classified 83% of the sample.

Additional evidence is available that cognitive dysfunction is not only strongly associated with the disorder in children and adolescents but also may be genetically based (Coolidge, Segal, Stewart, & Ellett, 2000; Coolidge, Thede, & Jang, 2004). In a study of 314 twins, ages 5–17 years, BPD as assessed by a DSM-IV based parental measure of personality disorders and executive function deficits was found to have a .77 heritability estimate and .52 correlations with executive function deficits (Coolidge et al., 2004).

Although the empirical evidence in children is supportive of the model proposed, the relationship between the child and adult disorder is poorly understood. Only one study was found that presents evidence for the validity of the adult diagnosis of BPD in children with BD (Bemporad, Smith, Hanson, & Cicchetti, 1982). One prospective study found that 19 out of 32 children diagnosed as borderline had an array of adult personality disorders but not BPD per se (Lofgren, Bemporad, King, Lindem, & O'Driscoll, 1991). Another study illustrated an association between the presence of major depression, disruptive behavior disorders, and/or substance abuse disorders in childhood or adolescence and the adult diagnosis of personality disorders (Ramlindt, Von Knorring, Von Knorring, & Ekselius, 2003). The authors concluded that Axis I conditions predispose to adult PD. However, PD may have been present during childhood and, undetected, influenced the expression of the Axis I condition. A method for diagnosing personality disorders in children is now available, and may further our understanding of this important area (see Coolidge, Thede, Stewart, & Segal, 2002). As with other neurodevelopmental disorders such as schizophrenia, it is probable that BD is only diagnosed in those children with a high degree of dysfunction. The disorder may not be detected until the developmental demands of young adulthood and/or adverse life events precipitate a significant disorganization in function and entry into treatment in the late teens and early 20s.

Clinical descriptions

The term “borderline” originated from the view that the disorder shared features of both psychosis and neurosis pointing towards a cognitive disturbance as a key feature of the disorder. However, it took 14 years from the first appearance of BPD in the DSM-III (American Psychiatric Association, 1980) for the cognitive features of paranoia and dissociation to be included in the diagnostic criteria for the disorder. This inclusion followed a series of empirical studies finding a high prevalence of stable cognitive disturbances in adults with BPD (Chopra & Beatson, 1986; George & Soloff, 1986; O'Connell, Cooper, Perry, & Hoke, 1989; Pope, Jonas, Hudson, Cohen, & Tohen, 1985). These disturbances have been characterized as unusual beliefs (e.g., superstitiousness, magical thinking, belief in a sixth sense, clairvoyance, or telepathy) dichotomous (black/white) thinking, overvalued ideas, unusual perceptions, severe dissociation, paranoia, and transient psychotic thought (Zanarini, Gunderson, & Frankenburg, 1990). Cognitive symptoms seem to occur as frequently as they do in schizotypal personality disorder and are a stable part of the disorder (Sternbach, Judd, Sabo, McGlashan, & Gunderson, 1992).

Neuropsychological studies in adults

The presence of cognitive problems suggested a potential link between BPD and underlying brain dysfunction, which led to a broad investigation into cognitive functioning assessed by neuropsychological measures. At least 18 neuropsychological studies of BPD have been published. These studies have examined inpatients and outpatients both off and on psychotropic medications, without schizophrenia, current psychosis, history of serious head injury, neurological disorder, low IQ, or current substance abuse in comparison to normative samples, healthy normal or clinical controls.

Fourteen of these studies have found cognitive deficits in BPD patients when compared to matched normal and/or clinical controls or normative data (Bazanis et al., 2002; Burgess 1990; Carpenter, Gold, & Fenton, 1993; Dinn et al., 2004; Dowson et al., 2004; Judd & Ruff, 1993; Lenzenweger, Clarkin, Fertuck, & Kernberg, 2004; Monarch, Saykin, & Flashman, 2004; O'Leary, Brouwers, Gardner, & Cowdry, 1991; Stevens, Burkhardt, Hautzinger, Schwarz, & Unckel, 2004; Swirsky–Sacchetti et al., 1993; Travers & King, 2005; Van Reekum, Conway, Gansler, White, & Bachman, 1993; Van Reekum et al., 1996). Four studies have reported negative findings, one without a control group (Cornelius et al., 1989), two compared to normal controls (Driessen et al., 2000; Kunert, Druecke, Sass, & Herpertz, 2003), and one compared to both a clinical and normal control group (Sprock, Rader, Kendall, & Yoder, 2000).

There are many challenges in the interpretation of neuropsychological findings. Most studies are not based in a comprehensive theory of brain functioning. Each test can measure multiple and different cognitive functions and domains that correspond to varying regions of the brain. For the purpose of discussion Luria's (1973) theory of brain functioning provides a basis for interpretation that can be applied across at least six of the controlled neuropsychological studies that used a similar comprehensive battery (see Table 1). Luria's theory was verified by a factor analytic study of a neuropsychological test battery similar to those listed in Table 1 (Baser, 1985; Judd, 1989).

Summary of significant findings in neuropsychological studies using a similar comprehensive battery according to Luria's theory of brain functioning

Luria (1973) proposed three functional units of the brain that are hierarchical in structure. Unit I regulates arousal, which includes measures of attention and vigilance and refers to a ready state primed for neural activity and necessary for rapid processing of information. Unit II includes mnestic and complex information processing tasks, which include measures of verbal and visuospatial processing, integration, learning and memory, and speed of processing. This involves the transformation of concrete perceptions into symbolic schemes necessary for encoding and storage of information. Unit III is responsible for goal directed intentional behaviors and includes executive function measures of abstraction, planning, and organization that entail the identification of a problem, investigation into the conditions of the problem, and analysis of its components.

Based on this analysis, BPD patients appear to have impairments primarily in Luria's Unit II consisting of mnestic and complex intellectual processes. These processes involve creating meaningful associations among stimulus cues and then forming these into multidimensional systems where elements are simultaneously synthesized. It also involves the rapid processing and complex integration of auditory, visual, and sensory information and their transformation into internal schemes, symbols, and abstract representations. All of these processes are central to the integration of information into memory systems, which appears to be impaired in patients with BPD.

Findings can also be clustered according to sensory modality and cognitive domain. Applying this approach to studies reported in Table 1, impairments occur as follows: (a) visuospatial modality, that is, Rey–Osterreith, block span, digit symbol, figural fluency, porteus mazes, block design, picture arrangement; (b) learning and memory, that is, Rey–Osterreith delayed recall, block span, Wechsler logical memory, and the California Verbal Learning Test; and (c) speed of processing, that is, digit symbol, trails A and B and the Continuous Performance Test. This approach highlights a deficit in nonverbal processing in areas of learning, memory, and speed of processing. These deficits could specifically interfere with rapid analysis and integration of incoming emotional and sensory information that might be incongruous as is typical of maltreating environments.

Another method of analysis of neuropsychological findings is to cluster impairments across various cognitive domains. In one study that applied a cognitive domain approach the BPD group was significantly impaired on the global impairment index and on seven of the nine domains (Monarch et al., 2004). Subjects were unimpaired on measures of abstraction and flexibility and fine manual motor skills. This finding suggests an impairment that may impact multiple areas of cognitive function.

In the largest controlled study to date, 42 BPD inpatients and outpatients were compared to normal comparison subjects on computerized neuropsychological tasks (Bazanis et al., 2002). In contrast to the studies reported in Table 1, BPD subjects were unimpaired on tasks of spatial and pattern recognition and visual memory, but did exhibit deficits on measures of visuospatial decision making and planning. The authors note that the impairment was partially due to longer deliberation time. Similarly, Dowson et al. (2004) compared BPD subjects to ADHD and normal controls and also found the BPD group was distinguished from the ADHD group by a longer deliberation time on a visuospatial decision making task. These findings add support for impairment in processing speed and/or allocation of resources on more complex visuospatial tasks.

These findings are further supported in a study of controlled or effortful processing that compared patients with BPD to normal matched controls on computerized measures of sustained attention, working memory, and executive functioning (Lenzenweger et al., 2004).

This study also explored the relationships between cognitive performance and the personality dimensions of positive and negative emotion and nonaffective constraint (control). The BPD group performed significantly worse than did the normal control group on the WCST, an executive functioning measure assessing abstraction and cognitive flexibility. These deficits were associated with decreased levels of control. State anxiety, depression, and positive or negative emotion levels were not significantly related to WCST performance, further supporting the specificity of the findings. The finding on the WCST parallels a similar finding in a study of children with BD described earlier (Paris et al., 1999). The WCST is also primarily a visuospatial measure supporting greater vulnerability in this area.

The likelihood that organic factors either genetic or acquired may underlie the cognitive deficits identified in patients with BPD has been further explored. Travers and King (2005) administered a neuropsychological battery to 80 BPD subjects divided into those with and without a history of significant organic factors. Organic factors included ADHD, learning disabilities, loss of consciousness, traumatic brain injury, and epilepsy. For the purposes of analysis the authors divided subjects into two groups: the nonorganic group and the organic group. A comparison of the two groups found significant differences on a measure of verbal fluency, two measures that assess speed and accuracy of visuospatial processing, and an attention factor. In comparison to a normative sample the organic group performed significantly worse on the majority of measures, whereas the nonorganic group performed significantly worse on some measures only. The positive findings in both groups were similar to those listed in Table 1. This study highlights the prevalence of organic factors in patients with BPD and supports an association between organic factors and poorer performance on neuropsychological tests. However, it also suggests that known organic factors do not fully explain the presence of cognitive impairment, as the nonorganic group still performed more poorly than did the normative group on multiple measures.

Four neuropsychological studies failed to find cognitive deficits in BPD patients. The Cornelius et al. (1989) study was hampered by the use of a brief screening battery and lacked a control group matched for age and education. The Driessen et al. (2000) study, administered only four neuropsychological measures known to be sensitive to the disorder. Although the BPD group performed more poorly on all four measures the authors concluded that the level of depression accounted for the findings. Depression is an associated feature of the disorder, and this raises questions as to the validity of these findings. Kunert et al. (2003) found no differences between subjects and controls on a comprehensive computer assisted battery of tests. However, they used a nonstandard diagnostic threshold for the categorical diagnosis of BPD (see Lenzenweger et al., 2004). Sprock et al. (2000) compared BPD patients to both a clinical (depressive disorders) and normal control group. There were no significant differences among groups on most measures. The BPD subjects were more impaired than normal controls on two tasks, but the depressed group had poorer performance than the two other groups on some measures.

Although the cumulative evidence supports the presence of cognitive impairment on BPD, both negative findings and variability within findings suggest that there may be subgroups of BPD patients on this dimension. Support for this possibility comes from the Swirsky–Sacchetti et al. (1993) study, which found large variability within their all female BPD group on most neuropsychological measures. A post hoc subgroup analysis found an “impaired” and “intact” group. Sprock et al. (2000) also conducted a subgroup analysis and found that a small subset of BPD patients (N = 6/18) exhibited poorer performance than depressive and normal controls on memory functions.

In summary, most of the published studies indicate that BPD can be characterized by cognitive impairments primarily in visuospatial learning, memory and speed of processing. Further, it appears that this impairment may represent a particular difficulty with effortful processing and/or allocation of cognitive resources. The findings represent subtler forms of cognitive impairment as the individual test scores are generally in the average (albeit lower) range and intelligence scores are generally in the average or above average range. It is unknown how BPD performance compares to other personality disordered groups, although one study did find a distinctly different pattern from antisocial personality disorder (Dinn et al., 2004).

Within BPD samples there appear to be heterogeneous findings that represent both varying degrees of severity and area of dysfunction. This is to be expected, because if there is a cognitive impairment that is unique for BPD, it is embedded in a normal variation in cognitive functioning and co-occurs with learning disabilities, ADHD, traumatic brain injury, epilepsy, and other organic factors.

Psychophysiological studies

Findings from psychophysiological studies provide additional information to support both the presence of brain dysfunction and a deficit in processing speed and allocation of resources. Studies using electroencephalograms suggest global brain dysfunction as evidenced by abnormal diffuse slow activity (Cowdry, Pickar, & Davies, 1985; De La Fuente, Tugendhaft, & Mavroudakis, 1998; Snyder & Pitts, 1984). One study used a backwards masking paradigm and found that BPD subjects were slower than controls in stimulus perception and had a slower motor response, suggesting that visual processing accuracy was reduced (Stevens et al., 2004). Further, in a delayed matching to sample paradigm in which visual, auditory, and crossmodal stimuli were presented, the BPD subjects' performance was inferior to controls on all tasks, suggesting a performance deficit independent of the sensory system involved.

The P300 ERP has been used to assess the presence of information processing deficits in a variety of psychiatric conditions. The relatively few studies of P300 in BPD have found reduced amplitudes and prolonged P300 latencies using a simple auditory oddball task (Blackwood, St. Clair, & Kutcher, 1986; Drake, Phillips, & Pakalnis, 1991; Kutcher, Blackwood, St. Clair, Gaskell, & Muir, 1987). Confirming and extending these findings, a recent study of adolescents with borderline symptomatology compared to normal controls found reduced P300 amplitudes on a visual task (Houston, Bauer, & Hesselbrock, 2004). Those with borderline symptomatology also exhibited a decrement of processing of all stimulus types on a modified version of the Stroop Color–Word Test, which may be accurate in detecting subtle cognitive impairments (Polich, Pollock, & Bloom, 1994). The authors concluded that their findings may indicate a generalized dysfunction involving attention or resource allocation. Of note, the P300 studies of adults are similar to those reported in children as described earlier (Lincoln et al., 1998).

Magnetic resonance imaging (MRI) and functional MRI (fMRI) studies

There have been at least four reports using imaging studies that suggest abnormalities in brain structure and function associated with BPD. One study (Lyoo, Han, & Cho, 1998), pursuing a hypothesis regarding executive dysfunction in BPD, found a significantly smaller frontal lobe volume in BPD subjects compared to healthy subjects. The other three studies found reduced volume in the hippocampus and/or amygdala, which are considered central to memory and to emotional processing.

Stein, Koverola, Hanna, Torchia, and McClarty (1997) found evidence for hippocampal abnormalities in a study of patients with posttraumatic stress disorder, some of whom also had BPD. Driessen et al. (2000) found a 16% reduction of the bilateral volumes of the hippocampus and 8% smaller volume of the amygdala in a sample of 21 female BPD patients. Schmahl, Vermetten, Elzinga, and Bremner (2003) compared 10 BPD subjects to 23 controls and also found smaller bilateral volumes of the amygdala (21.9%) and the hippocampus (13.1%) in the BPD group. The left and right amygdala had the greatest magnitude of effect.

Three fMRI studies have explored patterns of brain activation in BPD patients and confirm abnormalities in the amygdala. One study compared six patients with BPD without comorbid Axis I conditions and six age-matched controls while viewing 12 emotionally aversive slides compared to 12 neutral slides (Herpertz et al., 2001). The BPD group showed a greater activation pattern bilaterally in the amygdala when viewing slides inducing negative emotions.

Another study compared 15 BPD subjects to 15 normal controls and found greater levels of left amygdala activation in response to faces exhibiting various emotions (Donegan et al., 2003). A third study examined a small sample of patients with DSM-IV Axis II Cluster B disorders (predominantly BPD) on a task assessing response inhibition (Vollm et al., 2004). Although Cluster B patients achieved successful response inhibition, they activated a more extensive left hemisphere neuronal network whereas activation in normal controls was predominantly in the right hemisphere. The authors suggested that the more widespread activation in the patient group may indicate a need to recruit additional cortical resources to effectively inhibit responding.

In summary, the impairments identified through neuropsychological and psychophysiological studies appear to reflect a predominant visuospatial performance deficit in speed of processing, encoding and memory, and/or a problem with general resource allocation that impacts a range of cognitive functions including executive functions. Preliminary imaging studies point to brain dysfunction that expresses itself in the amygdala and hippocampus both central to memory processes. The amygdala has also been described as a rapid transient information processing pathway for emotionally relevant stimuli (LeDoux, 1996). These findings support the presence of impairment in cognitive processes that may influence the rapid encoding and storage of emotional information through the amygdala into memory systems.

Metacognitive monitoring

As depicted in Figure 1, a proposed further developmental outcome of dissociation and disorganized attachment is impaired metacognition in the interpersonal arena (see Main, 1991; Main & Goldwyn, 1998) Interpersonal metacognition corresponds to the level of formal operations in Piaget's theory and empathy is the hallmark of formal operations in the interpersonal arena (Lane & Schwartz, 1987). Hughlings Jackson hypothesized that the higher psychological functions (i.e., the sense of self, consciousness about self, and self-reflective function) are the most recent in evolutionary history, the most vulnerable to maturational disturbance, and the most experience dependent (see Meares, Stevenson, & Gordon, 1999). Fonagy, Target, and Gergely (2000) have proposed that disorganized attachment interferes with the development of reflective function which helps to integrate self–other representations.

Metacognitive monitoring refers to the ability to observe oneself while speaking, and to detect errors in reasoning or inconsistencies in one's narrative (i.e., to think about thinking; Flavell, 1979). Metacognitive knowledge refers to the recognition of an appearance–reality distinction that things may not be as they appear, and that appearances are never certain (Flavell, Flavell, & Green, 1983). It also refers to an awareness that the same things might appear differently to different persons, and that our thoughts vary from day to day about the same topic (Forguson & Gopnik, 1988). These operations and knowledge, which appear impaired in patients with BPD, are central to empathic ability and require a high degree of analytic, synthetic, and flexible thinking. They are hypothesized to be related to secure attachment and an integrated representational model of self and others (Main, 2000).

Impaired metacognition has been identified in adults with insecure, disorganized attachment as assessed by the Adult Attachment Interview (AAI; George, Kaplan, & Main, 1996; Hesse & Main, 2000). Disorganized attachment is scored on the AAI by an analysis of the person's narrative use of language and the ability to reflect upon significant interpersonal events while regulating emotion about these events and attending to the collaborative goals of the interview. Those with disorganized attachment are described as “incoherent speakers” related to their inability to provide a clear narrative about their personal history and to stay focused on the collaborative nature of the interview. Their narratives are replete with illogical thinking and lapses in reasoning about their attachment figures.

Dissociation in everyday life

Based on the model proposed here, dissociation would be a constant feature of the psychological life of those with BPD and impact the person with BPD on a continuum from mild to greater severity. In this hypothetical context the clinical manifestations of dissociation relate to the degree of difficulty integrating and processing emotionally arousing information. During periods of low stress individuals with BPD may feel mildly disconnected and unaware of their emotional life. However, they would be able to process most interpersonal information in an adaptive manner although still missing key elements. More stressful times would interfere to a greater degree with processing of arousing information, and would lead to feelings of depersonalization and derealization. Amnesia could occur in states of extreme stress, and represent a complete absence of cognitive processing of the arousing event with sense and emotional memories compartmentalized although capable of retrieval in psychotherapy.

Denial, projection, and splitting as manifestations of impaired metacognition

Kernberg (1967) first described the clinical manifestations of borderline personality organization in adults, and highlighted the patient's reliance on the defenses of denial, splitting, and projection. These defenses may be better understood as manifestations of inadequate analysis and processing of interpersonal information, impaired metacognitive monitoring, and fluctuating metacognitive knowledge as defined above.

Denial refers to the apparent lack of awareness of many aspects of the self, including the emotional impact of other's behaviors on oneself and of one's own behaviors on self and others. BPD patients are often unaware of their feelings of hurt and anger, and when asked, will look blank, minimize, or deny. They minimize or ignore the dangerousness of their self-damaging behaviors as though it were not happening to them. Although it is beyond the scope of this paper to elaborate all of the processes involved in denial, a significant aspect is an impaired ability to cognitively process emotional information as part of an analysis of an interpersonal event.

Similarly, splitting implies an inability to think dichotomously, that is, to entertain opposing thoughts about others and understand that they have competing motivations and separate states of mind and also to realize that thoughts and actions vary day to day. Significant others are experienced as all good or bad, based upon immediate experience rather than on a continuous and integrated experience of the person. Kernberg (1967) noted that, “Splitting interferes with the ability to synthesize which normally brings about … abstraction and integration of an object relationship” (p. 672), but it may more accurately be described as an outcome of the inability to integrate and abstract in the interpersonal arena.

Projection refers to the attribution of one's own thoughts and emotions to another. This reflects lack of awareness and confusion as to who is the owner of one's emotions as they have not been integrated into one's personal history. Projection incorporates elements of denial and splitting and contributes to a variety of cognitive distortions, paranoia, and paranoid states. Here, too, the individual is unable to observe and self-reflect, and include the self and interpersonal context in the analysis.

Unusual beliefs, cognitive distortions, and black and white thinking

Impaired metacognition can result in a variety of unusual beliefs, cognitive distortions, and rigid thinking patterns that impair social relationships. Patients with BPD often rely on magical thinking, telepathy, clairvoyance, and a “sixth sense” to understand interpersonal situations. This reflects an underlying difficulty in analyzing interpersonal interactions in an integrated fashion through use of empathy. Patients with BPD intuit part of the interpersonal picture but make the wrong interpretation and at a higher or lower level of abstraction than is needed. Their interpretation of interpersonal situations is distorted in that they cannot reflect upon their role, the context, and the other's point of view, but focus on their own immediate needs.

Another developmental outcome of impaired metacognition is all or nothing/black and white thinking. Patients with BPD are unable to see the middle ground or “grey” in a variety of situations. They rely on overvalued ideas as a means to guide their behavior. This inflexibility in their thinking impairs interpersonal problem solving and interferes with new learning.

Transient psychotic states

When in a state of extreme crisis and unable to obtain needed care from a significant other, the BPD patient's underlying disorganized attachment and impaired metacognition become most apparent. They are prone to brief losses of reality, transient hallucinatory, or illusory phenomena and paranoid states. These states represent a collapse of problem-solving strategies and a complete breakdown in the coherent processing of information.