Emotional Trauma and Childhood Amnesia (Reprinted from: the Journal of Consciousness & Emotion, 4 (2), 151-178, 2003), by Rhawn Joseph
Rhawn Gabriel Joseph, Ph.D.
Brain Research Laboratory
It was determined that among the general population childhood amnesia (CA) has its offset at around age 3.5 with females forming their first recallable memories earlier than males. In general first memories tend to be visual and positive rather than verbal or negative, and the period for CA is greater for verbal vs visual memories and for negative vs positive memories. Because negative, and repetitive stressful experiences disrupt memory and can injure memory centers such as the hippocampus and amygdala, and since adults who were traumatized or abused (TA) during childhood reportedly suffer memory disturbances, it was hypothesized that
those with a history of repetitive instances of early trauma, might suffer from a lengthier period of childhood amnesia and form their first recallable memories at a later age as compared to the
general population (GP). Because the non-motor regions of the right hemisphere matures earlier than the language dominant left hemisphere, and is dominant for visual and emotional memory, as well as the stress response, and as stress can injure the brain and disrupt memory, it was hypothesized that the traumatized group would demonstrate memory and intellectual disturbances associated with right hemisphere injury as based on WAIS-R, Wechsler Memory
Scale and facial-memory testing. All hypotheses were supported. Positive and visual memories are formed before negative and verbal memories. TA CA offset, on average, is at age 6.1 versus 3.5 for GPs, with TA females forming their first recallable memories at a later age than TA males. Thus the sex differences in CA is reversed. TA PIQ, short-term visual memory, and facial memory were significantly reduced. The results indicate, therefore, that childhood amnesia has its offset at around age 3.5 for the general population and 6.1 for those with a history of early trauma, and that repetitive instances of traumatic episodes significantly disrupt memory and neurological functioning such that individuals who were repeatedly traumatized
experience a period of childhood amnesia which is almost twice that of the general population.
Over the course of the last 100 years, it has been repeatedly established that most adults, adolescents, and older children, are unable to remember events from infancy or early childhood (e.g., Fitzgerald, 1991; Freud, 1905; Hall, 1899; Mullen, 1994; Potwin, 1901). This condition is referred to as "childhood amnesia."
Amnesia for childhood appears to have its offset at about age 3.5, on average, with females forming their first recallable memories earlier than males. In 4 separate studies it has been reported that on average, males form their first recallable memories at 3.64, 3.67, 3.67, and 3.8 years of age, whereas females recall their first memories formed at 3.23, 3.4, 3.5, and 3.6 years of age (Dudycha & Dudycha, 1937, 1941; Gordon, 1928; Waldvogel, 1948; see also Kilhstrom & Harackiewicz, 1982; Rubin, et al 1986).
However, although there are those who claim to recall events from before age 2, some individuals recall nothing from before age 10 (Joseph, 1998, 1999). It has been reported (Joseph, 1999), that those claiming no memory for their childhood were victims of repetitive instances of severe emotional or sexual trauma.
Disturbances of early memory have been frequently reported among adults who were traumatized or abused as children (Berliner, et al. , 2003; Briere & Conte, 1995; Courtois, 1995; Duggal & Sroufe, 1998; Elliot, 1997; Wessel, et al 2002). The more prolonged or repetitive the trauma, and the younger the victim, the greater is the degree of amnesia (Joseph, 1998, 1999).
Severe disturbances of memory, and related post traumatic stress disorders have also been reported for victims of childhood sexual abuse (Briere & Conte, 1995; Courtois, 1995; Duggal & Sroufe, 1998; Elliott, 1997; Williams, 1994), rape (Burnam, et al 1988; Christianson & Nilsson, 1984; Donaldson & Gardner, 1985), physical assault and injury (Kolko, 1996; Kuehn, 1974; Loughrey, Curran, & Bell, 1993; Peterson & Bell, 1996; Resnick,et al. 1993), front line combat (Blake, Cook, & Keane, 1992; Grinker & Spiegel, 1945; Kulka, et al. 1990; Southard, 1919), and natural disasters and related trauma (Escobar, et al. 1992; Fisher, 1982; Leopold & Dillon, 1963; Terr, 1994; Wagennar & Groenenweg, 1990).
Coupled with the evidence discussed above, it might be expected that traumatized children would experience an offset of childhood amnesia at a later age, depending on if the trauma was a single event or repetitive over an extended period of time.
For example, animals subjected to repeated or prolonged period of stress generally become amnesic and suffer injury to the hippocampus (Conrad, et al 1996; Diamond, et al. e,1994; Kirchbaum, et al 1996; Pavlides, Nivon, & McEwen, 2002; Pham, Nacher, Hof, McEwen, 2003), the hippocampus playing a major role in memory (Maguire, Vargha-Khadem, Mishkin, 2001; Morcom, et al 2003; Piefke, et al. 2003; Zeineh, Engel, Thompson, & Bookheimer, 2003).
By contrast, single, brief, traumatic or unpleasant episodes are more likely to be remembered, and emotional and memory centers, i.e. the amygdala and hippocampus, are less likely to be injured.
Usher and Neisser (1993) claim first memories are formed around age 2 for a single negative event (hospitalization). Unfortunately, the results from the Usher and Neisser (1993) study are based on college students who first heard family stories and/or had seen pictures of, and who earlier filled out a questionnaire documenting the negative events they were later supposed to remember during the experiment. That is, they were first asked about "hospitalizations" etc.
The validity and generalizability of the above results appear to be contradicted by another study by Neisser and Harsh (1992) who had subjects fill out a questionnaire regarding the Challenger space craft explosion. When questioned 32-34 months later, 75% could not recall filling out the questionnaire and many had forgotten considerable detail regarding the accident. According to Neisser and Harsch (1992), "As far as we can tell, the original memories are just gone."
In fact, memory for negative public events decreases with decreasing age, with young children showing the highest rate of forgetting (Warren & Swartwood, 1992; Winograd & Killinger (1983); a fuction, possibly, of age-related differences in frontal and temporal lobe memory structures (Morcom, Good, Frackowiak, & Rugg, 2003) and the immaturity of the neocortex and the interhemispheric fiber pathways, the corpus callosum (Joseph, 1982; Joseph, Gallagher, Holloway & Kahn, 1984).
For example, in examining memory for the deaths of President Kennedy and Robert Kennedy, Winograd and Killinger (1983) found a steep gradient of forgetting which became more profound with decreasing age. Those who were younger than 3 when these traumatic events occurred had no verbal recollections regarding information sources, context or associated events, and only approximately 50% of those who had been 4.5 years old or older could verbally recall the news and provide at least one verbal detail when questioned as adults.
Likewise, Warren and Swartwood (1992) in their memory study of the Challenger explosion, found that children age 5 were less accurate and more likely to delete features and forget details as compared to older children. In fact, those who had been the most emotionally upset suffered the greatest degree of memory loss.
Moreover, it has been repeatedly demonstrated in animal and human studies, that exceedingly negative events and highly arousing experiences disrupt memory and are forgotten to varying degrees (Christianson, 1992; Conrad et al., 1996; Diamond, et al.,1994; Kirchbaum, et al., 1996; Newcomber, et al 1994; Yerkes & Dodson, 1908); a consequence, presumably, of excessive amygdala and hippocampal activity or stress-induced injury to these memory structures (Lupien & McEwen, 1997, McEwen, 2002; Pavildes, et al., 2002; Pham et al., 2003, Sapolsky, 1996).
For example, hippocampal atrophy has been documented among adults sexually abused as children (Bremner, Randall, Southwick, Kyrstal, & Innis, 1995a; Stein, Hannah, Koverloa, & McClarty, 1995), and adults stressed from front line combat (Bremner, Randall, & Scott, 1995b; Gurvits, Shenton, Hikama, Ohta, Lasko, & Gilbertson, 1996).
Right and left hippocampal reductions have also been demonstrated in rats reared under deprived conditions (Diamond, 1985), and deprived rats also display significant disturbances in immediate and long term visual memory (Joseph, 1979, Joseph & Gallagher, 1980).
Moreover, the non-human amygdala (Cain, 1992; Henriksen, Bloom, & McCoy, 1978; Kraemer, 1992), the right amygdala in particular (Diamond, 1985) has been shown to become reduced in size, and/or to develop abnormal activity in response to severe stress including deprived rearing conditions. The amygdala plays a major role in emotional memory (Hall, et al 2001; Roozendaal, et al 2001), and when excessively aroused or highly stressed, memory is several disrupted (Gloor, 1997; Joseph, 1998, 1999; Kim, Lee, Han, &. Packard, 2001).
Likewise, Goodman and colleagues (Goodman, et al. 1991; Saywitz, Goodman, Nicholas, & Moan, 1991) found that painful experience and high levels of arousal (a painful vaginal and anal exam) interfere with memory functioning and actively "suppressed memory." Goodman, et al., (1991; Saywitz et al., 1991) report that children up to age 7 required cues and specific questions before they were able to recall the details, even following extensive questioning. "In free recall and demonstration... 78% and 83% failed to report it, respectively" (Saywitz et al., 1991, p. 686).
Again, the younger the victim and the more stressful and prolonged the trauma, the more pronounced might be the memory loss such that in some cases there results a complete dissociative amnesia. Consider, for example, the well publicized trauma of "Baby Jessica" McClure who at age 18-months was trapped in a narrow hole, 22 feet below the Earth's surface, for 58 terrifying and painful hours (Babineck, 1997). Although internationally televised and subject to one television movie and several books, and despite the skin grafts and amputation of one toe, 10-year old Jessica cannot remember anything of her ordeal (Babineck, 1997).
Tara Burke was kidnapped at age-3, held captive in a van for 10 months, and repeatedly sexually assaulted by two men; an ordeal filmed by her tormentors. Despite intense media scrutiny and a subsequent court trial, 18-year old Tara reports that the "memory has been erased" from her mind. "It's like a story that has happened to someone else" (Joseph, 1998).
Thus it appears that highly stressful negative and traumatic events which are repetitive and prolonged, can injure the hippocampus and amygdala and disrupt memory functioning to varying degrees.
One might suspect, therefore, that first recallable memories are more likely to be positive than negative, which in fact is the case as has been repeatedly reported for over 100 years (Colgrove, 1899; Dudycha & Dudycha,1933, 1941; Hall, 1899; Kihlstrom & Harackiewicz, 1982; Potwin, 1901; Waldvogel, 1948). Indeed, formal studies of memory for traumatic versus positive childhood experiences have shown that trauma memories are emotionally upsetting, less coherent, riddled with gaps, and lack sensory details (Berliner, Hyman, Thomas, Fitzgerald, 2003; Wessel, Merckelbach, & Dekkers, 2003).
It can also be predicted that, on average, long-term negative memories are established at a significantly later age than positive memories.
Likewise, it can be predicted that the period of childhood amnesia would be more extensive among those individuals who were repeatedly and severely traumatized at an early age. That is, as early experiences were negative, the length of childhood amnesia would therefore be greater, and first recallable memories would be formed at a later age, due to the deleterious effects of stress on memory and the brain.
Thus, one of the purposes of this study was to determine if adults who had been severely traumatized during childhood form their first recallable memories at a later age as compared to the general population. Yet another purpose was to determine if these individuals demonstrate cognitive, memory, and intellectual deficits associated with stress-related neurological injury to temporal or frontal memories centers, including the amygdala and hippocampus. According to studies performed on animals and humans, the greater the severity and/or if the trauma is repetitive the greater is the memory loss and the greater is the injury to the amygdala and hippocampus (Lupien & McEwen, 1997; McEwen, 2002; Sapolsky, 1996), structures implicated in memory.
Although negative vs positive experierences are important variables, it has also been argued that the offset of childhood amnesia is directly related to the child's growing vocabulary (Freud, 1905; Waldvogel, 1948). If true then it might be expected that first recallable memories would be predominantly verbal and include statements or words uttered by the child or family member, etc.
However, if first memories are visual and pictorial, this might indicate that the non-verbal visual, pictorial, emotional, and personal memory centers of the right hemisphere (Berckmoes & Stroobant, 2003; Cimino, et al. 1991; Evans et al., 1995; Joseph, 1988a,b; Levy & Heller, 1992; Wechsler, 1973) mature before the language-dominant left hemisphere which is dominant for non-emotional verbal memory (Fletcher & Henson, 2001; Morcom, et al., 2003).
Thus, one of the purposes of this study was to determine if negative memories are formed at a significantly later age than positive memories and if these memories are predominantly visual or verbal.
Prospective adult participants were approached between the hours of 11:30 A.M. and 1:30 P.M., at four major public locations in the city of San Jose, California (population 800,000), and were asked and told the following: "I am conducting a survey on early childhood memories. Could I take just a few minutes of your time to ask you a few questions about your earliest memories?"
A total of 132 adult males and females were solicited in this fashion. Thirty five initially declined to participate, giving reasons such as "this is my lunch hour," or "I've got just a few minutes to get back to work." However, three of these individuals later approached this examiner and indicated a willingness to participate. A total of 50 women and 50 men, aged 18-51, agreed to serve as participants in this survey.
All participants were asked the following questions in the following manner: "If you think back to when you were a child, what is your very first memory? Your earliest memory?
If participants did not understand the question, they were told the following: "For example, one fellow said that his very first memory was his mother buying him a rocking chair. That was as far back as he could remember. What is your earliest memory, and how old were you when this happened?"
As per age, if participants stated: "three" or "two and a half" or "between four and five," and so on, these responses were respectively scored as "3" or "2.5" or "3.5."
Once this information was provided, and in order to clarify and verify the visual-pictorial, verbal, or verbal-visual-pictorial nature of the memory, participants were asked: "Is this purely a visual picture-like memory, or does it have any words and sentences that you can clearly remember?"
If the first recallable memory did not have clear verbal components, participants were later asked: "What is the very first memory you can remember that has words or sentences in it? Something someone said, or that you said, and that you can clearly remember?"
Since 75% of first memories reported were clearly visual-pictorial, most participants were also asked for a first verbal memory.
If the participant stated, "I remember standing in my crib and I kept saying, ducky," or "I was hiding under my bed and my mother was calling me," these responses were scored as a verbal-visual memory as they had both visual and verbal components (verified by further questioning). Subjects were then asked for the age at which this event occurred.
If participant could not recall a childhood memory that contained explicit or vague words or sentences, this was noted. Hence, as there were no reported memories that were purely verbal, all responses were scored as visual-pictorial or verbal-visual.
After providing a first memory, participants were all asked: "Is this a positive memory, or a negative memory, or is it neutral?" This question was then repeated if participants were able to provide an additional verbal-visual memory.
Lastly, participants were asked their current age; and any questions they might have had about this survey were answered.
The average age of first visual memory for females was found to be 3.31 and for males, 3.75, whereas the average age for neutral, positive, and negative memories were found to be 2.93, 3.25, and 5.47 respectively.
Visual-Pictorial and Positive vs Neutral vs Negative Memories
A 2 x 3 (general by emotional value) ANOVA was performed for the variable age of first visual memory, and main effects were found for gender (F=4.13, df=1/94, p=0.45), and emotional value (F=35.52, df=2/94, p<000000001) such that females form their first memories at earlier age than males, and positive and neutral memories are formed at an earlier age than negative memories. This impression was confirmed via a Welch Modified Two-Sample t-Test (for unequal population variances) with a two-sided alternative hypothesis so as to compare emotional values of first memories. It was found that positive memories are formed significantly earlier than negative memories (t=4.88, df=18, p<0.0001), and neutral memories are formed significantly earlier than negative memories (t=5.21, df= 23, p<0.001). There was no significant difference between positive and neutral memories.
Two-sample tests for equality of proportion (with continuity correction) were performed to compare males and females in regard to positive, neutral or negative memories. No significant differences were found.
Verbal-Visual Memories and Positive vs Neutral vs Negative Memories
A 2 x 3 ANOVA (gender by emotional value) was performed for the variable age of first verbal memory, and main effects were found for gender (F=5.07, df=1/67, p=0.028), and emotional value (F=18.22, df = 2/67 p<000001).
However, a special features of this data is that 10 females and 17 males were unable to recall a first verbal-visual memory and thus its emotional significance. These gender differences in missing verbal memories, however, were not found to be significant (p=0.088; confidence intervals for males and females= -0.024, 1.00). Nevertheless, this problem of missing data was not initially treated as a regular missing value since that assumes these particular subjects, if provided more time, would not have been able to come up with a childhood verbal-visual memory (which, however, might be the case). Since discarding the missing data would likely provide a distorted mean, these "blanks" were replaced with a value of "10." When analyzed in this fashion, significant differences between males and females for age of first verbal-visual memory (t=2.07, df=98, p=0.0414) were found with a confidence interval for mean differences of 0.04 (females) vs 2.2 (males). Hence, females form their first recallable verbal-visual memories at an earlier age than males.
Because the above test is sensitive to the values inserted to replace missing data, a Wilcoxon rank-sum test was also performed as it is sensitive only to the ranks of the data values and not to particular values. This yielded a (two-sided) p-value of 0.0081, which again indicates that females form their first recallable verbal-visual memories at a significantly earlier age than males.
Two sample tests for equality of proportions (with continuity correction) were performed comparing males and females for age of first positive, negative, neutral memories and blanks (no memories). The only significant differences was in regard to neutral memories (p=0.034, confidence interval 0.028, 1.00), such that males had more neutral memories. There were no significant differences between age of first recallable verbal-visual positive vs negative vs neutral verbal-visual memories.
Male and Female: Visual-Pictorial vs Verbal-Visual Memories.
For females, on average, first recallable visual memories are formed earlier than verbal-visual memories; i.e. 3.31 vs 4.11 respectively. Likewise, for males, on average, first recallable visual memories are formed earlier than first recallable verbal-visual memories; i.e. 3.75 vs 4.56 respectively.
A Welch Modified Two-Sample t-tests were performed to determine if these differences are significant. Missing values were not replaced. It was found that visual memories are formed at a significantly earlier age than verbal memories for males (p=0.0001) and females (p=0.0089).
Positive vs Negative vs Neutral: Visual-Pictorial vs Verbal-Visual Memories.
A one-way ANOVA was performed comparing the emotional value of visual vs verbal-visual memories and significant differences were found (f=13.16, df=1/171, p=0.0004) such that visual memories are more positive than verbal memories which tended to be more neutral. No significant differences were found for age of first negative visual vs age of first negative verbal-visual memories, thus indicating that negative memories in general are formed at a later age than positive memories. However, since visual memories in general are more positive than verbal memories, visual memories are formed at a significantly earlier age.
Chi-squared tests for equality of proportions supports the ANOVA. Visual memories are more positive (X2=10.31, df=1, p=0.0013), and verbal memories are more neutral (x2=8.70, df=1, p=0.0032).
Male vs Female: Emotional Significance of First Memories To determine if there were significant sex differences in the emotional values assigned to first visual memories (i.e. positive, negative, neutral), two-sample tests for the equality of proportions (with continuity correction) were performed and no significant differences were found (p's=1, 0.59, and 0.51 respectively).
To determine if there are sex differences in the emotional significance attributed to verbal-visual memories two-sample tests for the equality of proportions (with continuity correction) were performed, and the only significant difference was for neutral memories (p=0.034). The confidence intervals for differences in proportions between males and females (0.028, 1.00) indicates that males are significantly more likely than females to assign neutral values to their first verbal-visual memories. Hence, the first verbal-visual memories of females are more emotional.
The average age of first memories among women and men from the general population is 3.31, and 3.75 years of age respectively (yielding an overall average of 3.53), with 2% claiming a first memory before the age of one and 2% having no memory of events before age 10.
These significant sex differences and averages are similar to most other studies of early memories (see introduction) and can thus be considered valid and may be generalized to the population as a whole.
As predicted, visual-pictorial memories (V-P) were formed at a significantly earlier age than verbal-visual memories (V-V). There were no first memories that were exclusively verbal. Not a single participant reported a V-V memory that was formed before a V-P memory. Thus, recallable V-P memories are formed at a significantly earlier age than V-V memories. In fact, 35% of the males and 10% of the females could not recall an early memory that contained explicit language or verbal-linguistic components, and of those which did, the majority were unable to explicitly recall what had been said.
As per males, 80% of first recallable memories were visual-pictorial and only 20% had verbal components (i.e. verbal-visual). As per females, 70% of first memories were visual-pictorial, and 30% were both visual and verbal. Thus, for both sexes, the first recallable V-V memories were formed, on average, nine months after the first V-P memories (female/male mean: V-P=3.3/3.7 vs V-V= 4.2/4.6) and these differences are significant.
In addition, positive and neutral first memories were formed on average, at a significantly earlier age than negative memories i.e. neutral=2.93, positive=3.25, negative-5.47. For males 52% and for females 54% of these first recollections were positive, which is also similar to other reports and can thus be considered valid and generalized to the population as a whole.
V-P memories were also more positive than V-V memories which were significantly more neutral and negative. These results are consistent with the hypothesis and other research findings which indicate that negative experiences have a detrimental effect on memory formation. Negative recallable memories are thus formed at a later age than positive or neutral memories.
Males were more likely than females to assign neutral values to the first recallable verbal memories. Hence, the verbal memories of females are significantly more emotional than those of males.
Since females were able to recall first memories that were formed at an earlier age as compared to males, and as female memories were more emotional, these results are also consistent with other reports which have demonstrated that female memories are more emotional and that their ability to recall emotional memories is superior to males (Canli, et al 2002; Walker, de Vries, & Trevarthen, 1987).
Overall, these results are consistent with the hypothesis that first recallable memories would be visual pictorial, and are supported by neuroanatomical and other maturational indices which demonstrate the non-motor regions of the right hemisphere mature at an earlier age than the language dominant left hemisphere (Chi, et al., 1977; Gilles et al., 1983; Joseph, 1982; Scheibel, 1993); the right cerebrum being dominant for most aspects of emotion (Berckmoes & Stroobant, 2003; Borod, 1992; Heilman & Bowers, 1996; Joseph, 1988a.b, 1996) as well as emotional, personal, and visual-pictorial memory functioning (Evans et al., 1995; Gainotti, et al 2003; Galletly, Clark, McFarlane, & Weber, 2001; Kimura, 1963; Levy & Heller, 1992; Piefke, et al., 2003).
Hence, first memories are visual rather than verbal may be due in part to earlier right hemisphere maturation and the slower developmental rate of the left hemisphere and related language functions.
It may also be a function of the immaturity of the corpus callosum, and thus an inability of the language centers of the left hemisphere to gain early access to the right hemisphere (Joseph, 1982, Joseph et al., 1984).
In addition, these results are consistent with the hypothesis and the results from other studies which indicate that negative experiences are not as easily recalled as positive experiences. Hence, recallable positive (and neutral) memories are formed at an earlier age than negative memories.
A restrospective review of patient files was conducted so as to identify those individuals who had undergone neuropsychological testing and who reported a history of frequent beatings, constant emotional abuse, or oral/anal/genital sexual penetration, and who described their childhoods as "miserable," "nightmarish" and so on. Because of these inclusion requirements, only the patient files of those individuals who had been referred to the Neurobehavioral Center (Brain Research Laboratory, San Jose, California) by the United States and California State Departments of Disability, for the purposes of neuropsychological assessment and disability evaluation, were included in this retrospective analysis.
These patients were applying for disability benefits due to alleged mental, emotional, psychological, or neurological disorders, or substances abuse. Almost all had been questioned regarding their early memories and were specifically asked to characterize their childhoods as either "good, happy, wonderful, miserable, horrible," etc. Almost all were asked if they had suffered sexual, physical, or emotional abuse, or other traumas. If they responded in the affirmative, they were questioned as to the nature, degree, and duration of the abuse. In addition, almost all were administered the (WAIS-R) Wechsler Adult Intelligence Scale-revised (Wechsler, 1981), and a battery of memory and visual-spatial tests, including the Wechsler Memory Scale (Russell, 1975; Wechsler, 1945) and a facial recognition memory test (Joseph, unpublished). This data (where available) was tabulated and included in the data analysis.
Approximately 1,300 patient files were examined. Of these, a total of 317 met the initial criterion for inclusion; i.e. history of abuse, neuropsychological testing, data on first memories. However, of these, 98 were excluded. Specifically, patients who reported a history of severe abuse, but who also reported a history of being moderately or severely learning disabled, or who appeared severely depressed, or who had a history of stroke, tumor, severe head injury with loss of consciousness for over 24 hours, degenerative neurological disorders or uncontrolled epilepsy, and those who were intoxicated, drugged with narcotics, acutely psychotic, mentally retarded, or obviously malingering when examined, and those who were not administered the WAIS-R, were excluded from this study.
Therefore, of the approximately 1,300 adults referred for disability evaluation, 219 met the final criterion for inclusion. Of these, 123 were adult males and 96 were adult females, with an age range of 18 to 63. All reported histories of severe and repeated instances of sexual, physical, and/or emotional abuse, and typically described their childhoods as "miserable," "nightmarish," "horrible," and so on. As per those individuals who claimed childhood abuse, but who also reported no memory before age 10 (that is, other than memories of "screaming," "yelling," and horrible emotional feelings), their reports of early abuse were judged to be authentic based on claims that the abuse continued throughout their teenage years.
For the purpose of additional and comparative data analysis, patients were assigned to one of three groups: Emotional abuse, physical abuse, sexual abuse. Although it is recognized that emotional abuse is a component of physical and sexual abuse, not all who are emotionally traumatized are physically or sexually assaulted.
First Memory Ratings
Age of first memory and emotional significance of first memories were noted in the records and often the first memory was described in a single sentence. For example, the following first memories had been respectively scored as "Negative" or "Positive": "Step dad kicking the shit out of me." "Playing with cousins. Singing." However, in some cases, the emotional (positive, negative, neutral) significance of the memory was not noted, and these were treated as missing data. There was no indication in the records if first memories were predominantly visual-pictorial, or verbal, or verbal-visual, and no attempt was been made to score memories in this regard. Hence, this data was not available.
All patients included in this retrospective study were administered the WAIS-R (Wechsler, 1981). This measure of intelligence consists of 6 verbal subtests and 5 non-verbal subtests, which are standardized so as to yield an age corrected verbal IQ (VIQ), performance IQ (PIQ) and full scale IQ.
The performance IQ (PIQ) of the Wechsler Adult Intelligence Scale-revised (WAIS-R) is associated with the functional integrity of the right hemisphere whereas the verbal IQ (VIA) is associated with the left (Chase, Fedio, Foster, Brooks, Di Chiro, & Mansi, 1984; Matarazzo, 1972).
The right hemisphere is dominant for most aspects of emotion, (Berckmoes & Stroobant, 2003; Borod, 1992; Joseph, 1988a.b, 1996) the stress response (Wittling, 1990; Wittling & Pfluger, 1990), as well as emotional, personal, and visual-pictorial memory functioning (Cimino et al., 1991; Evans et al., 1995; Gainotti, et al., 2003; Galletly, Clark, McFarlane, & Weber, 2003; Kimura, 1963; Levy & Heller, 1992; Piefke, et al., 2003; Wechsler, 1973). The right hemisphere is also differentially effected by prolonged stress and functions abnormally "during both the updating and use of working memory" (Galletly et al., 2001).
Thus it was predicted that the right hemisphere of those who had suffered severe emotional and other traumas would be more greatly affected than the left hemisphere and as compared to the general population, and this would include its ability to form first memories. Thus it was predicted that among severely abused populations the PIQ would be significantly lower than the VIQ and that right hemisphere memory functioning would be abnormal, due possibly to right cerebral injury secondary to chronic abuse and emotional stress during early chilhood.
For the purposes of this study, VIQs and PIQs were tabulated and compared to each other and to the minimum IQ that is considered "average," i.e. 90.
Wechsler Memory Scale
The verbal narrative and visual-pictorial portions of the Wechsler Memory scale were administered to all patients included in this retrospective study, and the results were scored and analyzed according to the Russell (1975) norms. Hence, an immediate and 30-minute delayed memory score for both visual and verbal memory were obtained and verbal and visual memory performance were compared and analyzed. However, as the raw verbal versus the raw visual memory scores are not equivalent, the Russell (1975) scaled scores were also obtained and analyzed so as to compare immediate and delayed verbal versus immediate and delayed visual memory.
Facial Recognition Memory Test
Difficulty or an inability to recognize faces is referred to as "prosopagnosia" and is associated with right temporal lobe and amygdala dysfunction (DeRenzi; 1985; Gainotti, et al., 2003; Jacobson, 1986; Tranel & Hyman, 1990).
The Facial Recognition Memory Test is an unpublished measure of facial memory which has been found to be sensitive to amygdala and temporal lobe injuries, the right temporal lobe/amygdala in particular (Joseph, unpublished). This test consists of 16 black and white photos of eight dark haired white males and eight dark haired while females. Patients are shown four pictures (two men and two women), one at a time, and are instructed to examine the faces closely as they will be required to remember and pick them out from a number of pictures which will be shown to them later. Each patient is allowed 10 seconds to examine each picture. Five minutes later, the patient is reminded of this test, and is then shown, one at a time, 16 photos which include the four photos examined earlier. Upon presentation of each photo, the patient is asked to indicate if this is one of the photos they had seen earlier. False positives and misses are scored separately but are then combined to yield total errors. Hence, a maximum of 16 errors are possible. However, as based on the results from a standardized pool of 25 neurologically intact "normal" adults, the average error rate has been found to be zero. As based on a sample of 50 neurologically impaired adults, an error rate of 1 is considered mildly disturbed, an error rate of 2-3 is considered moderately disturbed, and an error rate of 4 or more is considered severely disturbed and an indication of prosopagnosia. However, the Facial Recognition Memory Test was not administered to all patients included in this retrospective study. Missing scores were treated as missing data.
First Recallable Memories: Males vs Females, Abused vs Non-Abused
In order to determine if those with a history of abuse form their first recallable memories at a significantly later age than those from the general population, the raw data from experiment I was included in this analysis. In contrast to participants from the general population, in which the mean of first recallable memory is 3.53 years of age, a mean of 6.13 was found for patients reporting a history of severe childhood abuse. A one-way ANOVA (abused vs general population) was performed for age of first memory, and a significant main effect was found (F=89, df=1/317, p<0.0001), such that those with a history of severe abuse form their first recallable memories at a significantly later age than participants from the general population.
A Welch Modified Two-Sample t-Test with two-sided alternative (for populations with unequal variances) indicates that these differences are significant (t=11.5, p<0.001). A 95% confidence interval for the difference in mean ages between the two groups is 2.16, 3.05. Hence, on average, individuals with a history of severe abuse form the first recallable memories 2 to 3 years later than participants from the general population.
Emotional Significance of First Memories: Abused vs Non-Abused In order to determine if the first memories of those with a history of abuse are more negative, positive, or neutral as compared to non-abused samples, the raw data from experiment I was included in this analysis. Based on a Chi-squared analysis, it was found that the first memories of patients with a history of abuse are significantly more negative as compared to those from the general population. (x2=74, df=1, p<0.001), whereas the first memories of those from the general population are significantly more positive (x2=32, df=1, p<0.001) and more neutral (x2=18, df=1, p<0.001) than the first memories of those with a history of abuse.
Two-sample t-tests for the equality of proportions (with continuity correction) were performed for positive, negative, and neutral memories) and no significant differences were found for males or females.
Two-sample t-tests for the equality of proportions (with continuity correction) were performed in order to compare those who had been physically vs sexually vs emotionally abused and no significant differences were found in regard to the emotional value of first memories.
Male vs Female and Emotional vs Physical vs Sexual Abuse and First Memories
A 3 x 2 ANOVA was performed to determine if the age of first recallable memory differed depending on the type of abuse experienced and if the victim were male or female. Main effects were found for gender (F=5.10, df=1/213, p=0.025) but not for type of abuse experienced. Specifically, it was found that abused males formed their first recallable memories at an earlier age on average, as compared to females; i.e. 5.85 vs 6.48 (F=5.10, df=1/213, p=0.025). However, further analysis with the Welch Modified T-test indicates these sex differences were not highly significant (t=1.75, p=0.0816), due to the high variances within the sample, especially for females.
Wechsler Verbal and Visual Memories
Two-way ANOVAs were performed for the immediate and delayed verbal and visual memory scores as compared to the Russell (1975) norms and the "scaled" scores for "average" performance. Histograms and QQ-plots for these variables indicate a strong deviation from normality. However, no significant differences were found for abused males or females across the measures of immediate and delayed verbal memory as compared to the Russell norms.
Abused males and females perform significantly worse on delayed measures of visual memory as compared to the Russell (1975) norms. Thus there is a significant degree of forgetting after a 30 minute delay as demonstrated by comparing ranked scores. A Wilcoxon rank-sum tests for both sexes indicates a significant degree of forgetting for long term (30-minute delayed) visual memories: males (p=0.026), females (p=0.017), and overall (p=0.005).
Wechsler Verbal and Visual Memories: Abused Males vs Abused Females
Separate 3 x 2 (abuse type by gender) ANOVAs were performed for the variables, immediate and delayed verbal memory, and immediate and delayed visual memory, and significant main effects were found.
The performance of abused females across the measures of visual memory are significantly worse as compared to abused males, i.e. immediate visual memory (F=7.84, df=1/213, p=0.006), and delayed memory (F=6.09, df=1/213, p=0.014). Standard Two-sample t-Test confirms that females perform significantly more poorly as compared to males on measures of immediate (p=0.0382) and delayed visual-pictorial memory (p=0.0384).
Facial Recognition Memory
As based on a normative sample of 25 neurologically intact adults in which the average error rate is -0-, it was found, using a Standard One-sample t-test that those with a history of abuse perform significantly worse and thus show a significant degree of forgetting for facial stimuli (p<0.01). A 95% confidence interval for the mean error rate is 1.12, 1.65. No significant sex differences were found, although females, on average, made more errors (x=1.54) than males (x=1.25).
In addition, a 2 x 3 ANOVA was performed to determine if those who had been physically vs sexually vs emotionally abused differed on their performance on this test and no significant differences were found.
Verbal and Performance IQs
Separate two-way ANOVAs (gender by type of abuse) were performed comparing the VIQ and PIQ, and no main effects were found for gender. As detailed below, the PIQ was found to be significantly lower than the VIQs (p<0.05), and main effects were found for type of abuse on both the VIQ (F=3.88, df=2/213, p=0.022) and PIQ (F=3.17, df=2/213, p=0.04).
A Welch Modified two-sample t-Test, demonstrated that those who had been physically abused have a significantly lower VIQ compared to those who had been sexually or emotionally abused (p<0.05), and a significantly lower VIQ compared to those who had been emotionally abused (p<0.05). Conversely, those who had been sexually abused were found to have a significantly higher VIQ and PIQ compared to those who had been physically or emotionally abused (p<0.05). There were no significant differences between the physically and emotionally abused group on the PIQ.
A Welch Modified two-sample t-Test also demonstrated that when comparing VIQs to PIQs, that when considered as a group, those with a history of abuse obtain a significantly lower PIQ as compared to the VIQ with means for males/females: VIQ=93.04/89.69, PIQ=88.09/83.14. The 95% confidence intervals for the overall differences in VIQ and PIQ are 2.76, 7.55, and for females= 2.87, 9.84; and males= 0.95, 7.49.
Specifically, among this severely abused and traumatized sample, the average age of first memory for women is 6.5, and for men, 5.7, compared with 3.3, and 3.7 years, respectively for women and men from the general population. Moreover, 11.5% of those with a history of severe and chronic abuse claim no memory before age 10, which again is in contrast to the general population in which only 2% claimed no memories from before this age. Although 7.2% of those reporting abuse have first memories that were formed by age 2, 13% of those from the general population claim to recall memories from this age or before.
Hence, although there are exceptions, clearly, those who claim to have suffered severe and repeated instances of childhood sexual, physical, and emotional abuse and who experienced horrible childhoods, demonstrate, on average, a much lengthier amnesia for childhood which in some cases may encompass the entire first decade.
In addition, significant differences were found when comparing the VIQ to the PIQ and to established WAIS-r norms, such that the PIQ was significantly lower for abused males and females. Although it is recognized that these individuals may not have been motivated to do their best, as they were applying for disability benefits, the possibility of malingering cannot explain why the PIQ is more depressed than the VIQ, and why the VIQ does not significantly differ from established norms whereas the PIQ does, and why this is true for both men and women. Although it is true that individuals who are severely or psychotically depressed also display reduced PIQs (Dean, Ray & Seretny, 1987), severely depressed patients were excluded from this study.
Likewise, the abuse sample demonstrated significant disturbances of visual, pictorial and facial memory storage and retrieval, which is also consistent with a lower PIQ vs VIQ and right hemisphere dysfunction.
As is now well established, non-verbal, visual-pictorial memory and intellectual functioning are associated with the right hemisphere (Berckmoes & Stroobant, 2003; Chase et al., 1984; Joseph, 1988a,b; Matarazzo, 1972), and the right temporal lobe and hippocampus and amygdala (Andersen, 1978; Evans et al., 1995; DeRenzi; 1986; Jacobson, 1986; Kimura, 1963; Tranel & Hyman, 1990). Moreover, facial recognition, and deficits in facial recognition and visual memory are associated with the damage to the amygdala, hippocampus, and inferior temporal lobe, and the right amygdala, right hippocampus, and right temporal lobe in particular (Evans et al., 1995; DeRenzi; 1986; Gainotti et al., 2003; Gloor, 1997, Jacobson, 1986; Kimura, 1963; Tranel & Hyman, 1990).
Hence, the results from experiment II suggest that the right hemisphere and right temporal lobe memory structures may be more severely effected by childhood stress.
These results are also consistent with the findings from a number of independent laboratories regarding the deleterious effects of stress on the brain, the hippocampus and amygdala in particular, thus resulting in significant disturbances of memory (Lupien & McEwen, 1997; McEwen, 2002; Pavlides, et al., 2002, Pham et al., 2003; Sapolsky, 1996). In fact, adults abused as children (Bremner et al., 1995a; Stein, et al., 1995), and adults stressed from front-like combat (Bremner et al., 1995b; Gurvits et al., 1996), also demonstrate hippocampal atrophy.
Indeed, it has been repeatedly experimentally demonstrated that highly arousing conditions, or repetitive and prolonged episodes of extreme stress, can injure the hippocampus and amygdala which may suffer profound dendritic death and atrophy, or develop seizure-like activity (Cain, 1992; Gahwiler 1983; Goelet & Kandel, 1986; Henriksen et al., 1978; Kraemer, 1992; Lupien & McEwen, 1997; McEwen, 2002; Pavlides, et al., 2002, Pham et al., 2003; Sapolsky, 1996; Uno et al., 1989). Moreover, it has been demonstrated that severe stress, including the secretion of stress-hormones and neurotransmitters such as cortisol, significantly impacts and disrupts learning and memory (Diamond et al., 1994; Kovacs, et al 1976; Lupien & McEwen, 1997; Shors, et al, 1989; Spoont, 1992), including memory retrieval (de Quevain, Roozendall, & McGaugh, 1998) recognition memory (Wolkowitz, Reus, Weingartner, Thompson, Breier, & Doran, 1990; Wolkowitz, Weignartner, Rubinow, Jimerson, Kling, & Berretini, 1993) and the ability to discriminate between relevant and irrelevant stimuli (Lupien & McEwen, 1997).
Reductions in the size of the amygdala, and hippocampal dendritic death and atrophy, have also been induced by social deprivation and impoverished rearing conditions (Diamond, 1985; Greenough, et al 1978); conditions which are exceedingly stressful and which permanently disrupt new learning and long term memory, with females being more significantly effected than males (Joseph, 1979; Joseph & Gallagher, 1980) as was the case in this study.
Similar to most studies, first recollection among men and women from the general population are formed around 3.5 years of age on average. This contrasts with an average of 6.1 years for adults who report a childhood history of repeated episodes of severe sexual, physical or emotional abuse.
Specifically, among women from the general population, the average age of first recallable memory is 3.3, and for men, 3.7 years of age. By contrast, among women with a childhood history of trauma or abuse, the average age of first memory is 6.5, and for men, 5.7.
Hence, although there are exceptions which are likely due to individual and gender differences and/or differences in the degree and nature of the trauma, those who have suffered severe and repetitive instances of sexual, physical, and emotional abuse during early childhood, experience a much lengthier amnesia for early childhood as compared to the general population. These individuals also display significant long term non-verbal intellectual and memory deficits associated with the right hemisphere, hippocampus, and amygdala.
Similar to the findings of numerous investigators, the first memories of participants from the general population are generally positive (experiment I). Moreover, negative first memories are formed at a significantly later age than positive experiences. The negative effect of unpleasant experiences on memory formation also explains why adults with a history of severe and repetitive negative experiences form their first recallable memories at a significantly later age as compared to those from the general population.
As demonstrated in experiment I, first recallable memories are also visual-pictorial and are formed at a significantly earlier age than verbal memories. As visual-pictorial memory is associated with the right hemisphere and the right temporal lobe and right amygdala and hippocampus, these results are consistent with those developmental studies which indicate that the non-motor regions of the right hemisphere mature at an earlier age than their left sided counterparts (Chi, et al., 1977; Gilles et al., 1983; Joseph, 1982; Scheibel, 1993; Thatcher, 1992). Again, an earlier maturation of the right hemisphere, coupled with an impoverished vocabulary, would explain why 70% to 80% of first recallable memories are visual-pictorial (right hemispheric) and only nine months later (on average) come to include verbal components.
Because the right hemisphere is also dominant for most aspects of emotion, including the formation and recall of emotional and personal memories, and as this structure is also implicated in the stress response including cortisol secretion, the right hemisphere also appears to be more at risk for stress-induced injury. Indeed, stress-induced right hemisphere injury would explain why adults with a history of severe, prolonged, and repetitive abuse, not only form their first recallable memories at a later age on average, but why the PIQ would be significantly lower than the VIQ and as compared to norms from the general population. This would also explain why short-term (30-minute delayed) visual-pictorial and facial memory are significantly disrupted (especially among women).
On average, most adults are able to recall positive visual-pictorial memories that were formed around age 3.5 but are amnesic for everything before that age. Long-term verbal memories are formed on average, around 9 months later than visual memories. Likewise, negative memories are formed at a significantly later age than positive or neutral memories. Those subjected to repeated instances of severe sexual, physical or emotional abuse, display a period of childhood amnesia which is twice that of those from the general population, i.e. 6.1 vs 3.5. Thus negative experience do not promote memory formation but contribute to an extended period of childhood amnesia.
Age of First Recallable Memories
Adults Who Were Repeatedly Abused During Childhood. (N=219)
Age of First Recallable Memories
Adult Females (N=96) vs Males (N=123)
Who Were Repeatedly Abused During Childhood.
Age of First Recallable Memories
Adults Who Were Repeatedly
1-Sexually Abused, 2-Emotionally Abused, 3- Physically Abused
During Childhood (N=219).