Traumatic brain injury screening and neuropsychological functioning in women who experience intimate partner violence

Abstract

Objective: The potential for traumatic brain injury (TBI) to occur as the result of intimate partner violence (IPV) has received increased interest in recent years. This study sought to investigate the possible occurrence of TBI in a group of women who survived IPV and to measure the specific profile of cognitive deficits using standardized neuropsychological measures. Method: A comprehensive questionnaire about abuse history; neuropsychological measures of attention, memory and executive functioning; and measures of depression, anxiety and post-traumatic stress disorder were given to women who were IPV survivors, women who were sexual assault (SA) survivors, and a comparison group of women who did not experience IPV or SA. Results: Overall, rates of potential TBI, as measured by the HELPS brain injury screening tool, were high and consistent with previous studies. Consistent with potential TBI, lower scores were demonstrated on measures of memory and executive functioning compared to survivors of SA or those not exposed to violence. Importantly, significant differences on measures of memory and executive functioning remained, after controlling for measures of emotion. Of note, cognitive changes were highest among women who experienced non-fatal strangulation (NFS) compared to IPV survivors who did not. Conclusions: Rates of TBI may be high in women who survive IPV, especially those who survive strangulation. Better screening measures and appropriate interventions are needed as well as larger studies that look at social factors associated with IPV.

Keywords:

• Traumatic brain injury
• intimate partner violence
• domestic violence
• cognitive deficits
• memory
• executive functions
• non-fatal strangulation

Intimate partner violence (IPV) can include stalking, sexual violence, physical violence and psychological aggression committed by one intimate partner against the other, and the majority of victims experience more than one type (Krebs et al., 2011). IPV affects all genders; however, women are victimized more often and sustain more severe injuries; and frequently violence against women by men is part of a broader pattern of assertion of control that may be based in gender inequalities (Reed, 2008).

For the purpose of this study we investigated physical violence due to IPV. Physical violence in IPV can include hitting, slapping, kicking, punching, pushing and choking, or objects thrown or used to hit the victim (e.g. Cantos et al., 1994; O’Doherty et al., 2015). The most common areas of the body injured in IPV are the head and neck from a direct blow to the head (Corrigan et al., 2003; Tam et al., 2010) and unwitnessed head and neck injuries are significant markers for IPV (Wu et al., 2010). Therefore, it is worthwhile to investigate whether significant, possibly, repeated, injuries to the head could result in traumatic brain injury (TBI).

In addition to blows to the head, brain trauma can also be the result of nonfatal strangulation (NFS) (Nemeth et al., 2019). Strangulation is a less commonly addressed mechanism of neuropsychological impact after IPV but the reported prevalence is high (Shah et al., 2004) in this population. Recent publications have reported NFS rates in IPV survivors to be 83% (Nemeth et al., 2019) and 89% (Strack & Gwinn, 2011), despite it being only about 1% in the general population surveyed (e.g. (Sorenson et al., 2014). Several different categories of strangulation exist: hanging; ligature strangulation; manual strangulation and positional strangulation (Monahan et al., 2019). All of these categories involve mechanically cutting off the airway which can lead to subsequent neurological insult unless interrupted.

The mechanism by which strangulation leads to hypoxic-ischemic brain damage is two-fold. First, there is the likelihood of oxygen and glucose deprivation. This hypoxia can cause impairment in brain function (e.g. Lu-Emerson & Khot, 2010). Additionally, this leads to excitotoxicity from glutamate release, which affects the hippocampal formation first and then can include the thalamus and cortex (e.g. Schurr & Rigor, 1992). Therefore, there is a possibility that survivors of NFS from IPV could experience cognitive deficits, especially in the areas of memory and learning (e.g. Anderson & Arciniegas, 2010; Bichard et al., 2022).

Given reports of blows to the head and NFS, some studies have measured complaints of cognitive changes after IPV. Although not widely studied, survivors of IPV have reported cognitive deficits such as memory loss, mental fatigue, and confusion (Zieman et al., 2017), difficulty following directions, difficulty retaining information, difficulty concentrating, inability to initiate self-directed behavior, and difficulty with abstract thinking (Monahan & O’Leary, 1999) in addition to depression, anxiety, irritability, and symptoms of post-traumatic stress disorder (PTSD). While it is difficult to tease out the cause of cognitive deficits as being directly related to brain trauma versus secondary to disorders such as depression and PTSD, there is evidence to suggest that both contribute (Monahan & O’Leary, 1999; Lifshitz et al., 2019).

In fact, Twamley et al. (2009) specifically measured cognitive deficits related to just the PTSD following IPV in individuals who did not have severe brain trauma. While they did not screen out mild TBI (mTBI) from their participants, they did exclude anyone who reported a loss of consciousness (LOC) of longer than 10 min or requiring hospitalization of more than 24 h. Participants were given a comprehensive battery of neuropsychological measures including processing speed, visuoconstruction, visual memory, timed verbal fluency, inhibition, reasoning, and switching. Differences between those with PTSD and the comparison group were found only on timed tasks and not on the other measures of cognition. Specifically, the only significant differences between the groups were on the Delis-Kaplan Executive Function System (D-KEFS) Trail Making Letter Sequencing and Motor Speed and on the D-KEFS Letter Fluency and Design Fluency. The authors suggest that cognitive resources may be being diverted to coping with psychological distress and to unpleasant internal experiences, leading to reduced capacity for speed of performance. Thus, these data lend support to the idea that deficits beyond the kinds of timed tasks reported in their study may not be explained by PTSD only.

Most studies to date on IPV and TBI have focused on documenting the occurrence and rates of TBI among IPV survivors (e.g., Gagnon & DePrince, 2017). One early study (Corrigan et al., 2003) reported that 30% of women surveyed reported a period of loss of consciousness, and 67% of all respondents had residual neurobehavioral sequelae that were potentially brain injury related. More recently, rates have been reported to be 38% when simply asking if the woman had been struck in the head or strangled to unconsciousness (Cimino et al., 2019); 50% using the HELPS screening tool (Hunnicutt et al., 2019) or 74%-100% using the American College of Rehabilitation Medicine (ACRM) definition of mild traumatic brain injury (Valera & Berenbaum, 2003; Valera & Kucyi, 2017). This has been investigated in a range of populations including women in the military (Amoroso & Iverson, 2017; Gerber et al., 2014; Iverson & Pogoda, 2015).

Neuroimaging techniques have been used to determine if measurable changes in brain anatomy or physiology could be detected in survivors of IPV (Fennema-Notestine et al., 2002; Flegar et al., 2011; Valera & Kucyi, 2017; Valera et al., 2019). Differences between those with a history of IPV and those without have included reduced fractional anisotropy in the body of the corpus callosum (Flegar et al., 2011). A negative correlation between “brain injury score” based on self-report of prior injury and fractional anisotropy was reported in regions of the posterior and superior corona radiata (Valera et al., 2019). Individuals with IPV have also demonstrated altered functional connectivity in a group of regions that are specific to cognitive/emotional control including the anterior cingulate, left amygdala and middle temporal gyrus (Roos et al., 2017). Severity of TBI was shown to be negatively correlated with intrinsic functional connectivity between the right anterior insula and posterior cingulate cortex/precuneus. No differences were observed in hippocampal volume (Fennema-Notestine et al., 2002; Flegar et al., 2011) but differences between those who had survived IPV and those who had not were reported as smaller frontal and occipital gray matter volume (Fennema-Notestine et al., 2002). Given these findings, difficulty with executive functioning, memory, and learning is not surprising (Valera et al., 2019).

It is important to note that the majority of IPV survivors are women (Reed, 2008). There is some evidence that women experience different symptoms and different recovery trajectories after TBI than men. For example, women may be more prone to receiving psychiatric diagnoses after TBI (Fann et al., 2004) and women report greater numbers of neurobehavioral outcomes such as somatic, vestibular, and psychiatric symptoms (Farace & Alves, 2000). There is also evidence to suggest that the rate of recovery after TBI is longer for women than men (Bazarian et al., 2010). Of course it is unknown what the interaction of morphology (such as neck length), hormonal differences, and differences in the responses of health care professionals is that may contribute to differences in reported symptoms and medical diagnoses.

Valera and Berenbaum (2003) and Stein et al., 2002, used neuropsychological measures to investigate cognitive impairments. This included the Trail Making Test, Digit Span, California Verbal Learning Task (CVLT), and Ruff Figural Fluency Test. These authors created a “brain injury score” based on self-report of severity, recency, and number of brain injuries using the ACRM definition of mTBI, which includes blows to the head, the head striking an object, and acceleration/deceleration injuries but excludes anoxia (1993). They did not include women who had a recent mTBI from other causes but they did not define how recent or indicate if there were women who had mTBI in the past from other causes. There was no healthy comparison group so z-scores based on published normative data for each measure were utilized. Of the cognitive measures administered, only the CVLT and Trail Making were significantly correlated with the “brain injury score”. In a later study Valera and Kucyi (2017) used the CVLT and Trail Making Test Part B to investigate a relationship between performance on these tests and resting state functional connectivity. Stein et al. (2002) also administered a series of neurospsychological measures and determined that individuals who had experienced IPV performed more poorly than healthy controls only on measures of working memory, executive functioning and visuoconstruction. Those IPV survivors with PTSD show poorer performance then healthy controls on a test of set-shifting (Trail Making Part B).

Recently, one study reported data to support the hypothesis that strangulation can lead to cognitive deficits in IPV survivors (Valera et al., 2022). This study examined a brief series of cognitive measures (CVLT, Trail Making Test, Digit Span, and Ruff Fluency) but did not include any healthy comparison participants and relied on published normative data which can be limited, especially in terms of race and ethnicity. They reported a relationship between history of strangulation and performance on the CVLT as well as measures of depression and PTSD. This is the only study, to our knowledge, to investigate neuropsychological effects of strangulation in this population.

Therefore, more data are needed to understand the nature of cognitive impairments in women who have experienced IPV which may differ from standard studies of TBI incurred in motor vehicle accidents or sports injuries. It is important to compare those who experience potential brain injury to those who experienced emotional and psychological trauma to begin to tease out the effects of emotional trauma versus physical trauma. Finally, it is critical to examine women who experience blows to the head as compared to those who experience strangulation to understand if the severity and type of impairment differs between these groups. In the current study women who had experienced IPV were compared to women who had never experienced any violence. In addition, as a control for post-traumatic stress symptoms, a comparison group of women who had experienced sexual assault (SA) but no blow to the head or NFS was included. The literature on SA survivors suggests that there are higher rates of depression and anxiety (e.g. Quidé et al., 2018; Thurston et al., 2019) with mild to no cognitive deficits (e.g. Quidé et al., 2018). There have been some studies to suggest more significant deficits in cognitive functioning (e.g. Raskin, 1997) however, these studies involved childhood sexual assault survivors.

The current study hypothesized that 1) women who experienced IPV would show deficits in attention, memory, and executive functioning compared to women who did not experience IPV, 2) 3) Women who survived NFS would show greater impairments in learning and memory than women who survived IPV but not NFS, 3) Women who survived IPV will show evidence of likely TBI on the HELPS screening measure and women who survived SA will not, and 4) there will be a relationship between cognitive performance in the IPV group and self-reported mood and stress

Methods

Participants

IPV survivor group

Fifty IPV survivors were recruited through domestic violence shelters, support groups, and hospitals in the Hartford, CT area. Inclusion criteria: participants had to identify as women, have experienced at least one prior incidence of IPV (self-reported by questionnaire as physical, sexual, or emotional abuse by an intimate partner), be between the ages of 18 and 65 years, be fluent in English, and have no other prior neurological or psychological conditions, including TBI from other causes.

No violence (NV) group

Fifty participants who had not experienced IPV were recruited from through online advertisements. Inclusion criteria were the same as the IPV group, except the healthy participants must not have experienced any incidence of DV or IPV or other assault. Racial and ethnic demographics were similar to those who survived IPV.

Sexual assault (SA) group

Thirty-five participants who experienced sexual assault as adults but no history of blow to the head or strangulation, TBI or other neurologic diagnosis, were recruited from online support groups and women’s resource centers for a prior study.

All participants received $20/hour for their time. The study received approval from the Trinity College Institutional Review Board.

Materials

IPV participants and participants with no history of violence (NV) were given a questionnaire about their social and medical history, measures of cognitive functioning, emotion and mood, PTSD, and quality of life, described below. All clinical assessment measures were given according to standard testing procedures given in the test manuals.

The IPV Health History Questionnaire (IHHQ) included questions from the HELPS to determine history of possible brain injury (Picard et al., 1991). The HELPS is a commonly used screening tool with questions related to each letter of the acronym: H (have you ever hit your head or been hit in the head), E (have you been to the emergency room, hospital or doctor’s office because of an injury to your head), L (did you ever lose consciousness or become dazed or confused because of an injury to the head), P (do you experience any problems since injuring your head such as mood changes, trouble with concentration, or trouble getting things done) and S do you experience sickness since injuring your head such as dizziness, headaches, or changes in vision, smell or taste). A score of three or more on the HELPS is considered indicative of a potential brain injury. The IHHQ also had embedded the questions from the Rivermead Post-Concussion Symptoms Questionnaire (King et al., 1999). Finally, it collected data on education, employment, marital history, medications, drug and alcohol use, family health history, neurological disorders, and past accidents. Most importantly, the survey addresses IPV, determining when, for how long, with how many partners the IPV occurred.

Cognitive assessment included measures of attention, memory and executive functioning since these are the areas of cognitive functioning most often impacted by TBI (e.g. Prince & Bruhns, 2017; Raskin et al., 2014). Attention was assessed with the Brief Test of Attention (BTA) (Schretlen et al., 1996). Memory was assessed with the Hopkins Verbal Learning Test (HVLT) (Benedict & Brandt, 2007). Executive functioning was assessed via verbal fluency measures, which were the Animal Naming Test of the Boston Diagnostic Aphasia Examination (BDAE) (Goodglass et al., 2001) and the Controlled Oral Word Association Test (COWAT) (Benton, 1967); The Trail-Making Test (Reitan, 1958); and the Stroop Color-Word Interference test (45 s time limit) (Golden & Freshwater, 2002). The American National Adult Reading Test (Grober & Sliwinski, 1991) was administered as an estimate of premorbid intelligence. For each of these tests we characterized a score as impaired if it fell two standard deviations or more below the published standard sample normative data with any corrections for age, education, or other demographics that were included with the normative data. These published data sets were BTA (Schretlen, 1997), HVLT (Benedict & Brandt, 2007), Animal Naming (Goodglass et al., 2001), COWAT (Loonstra et al., 2001), Trail Making Test (Strauss et al., 2006) and Stroop (Golden & Freshwater, 2002).

The measures of mood were the Beck Anxiety Inventory (BAI) (Beck et al., 1988), and the Beck Depression Inventory-II (BDI-II) (Beck, Steer, & Brown, 1996). Both the BAI and the BDI measures are 21 item self-report rating inventories. For the Beck Depression Inventory, a score over 30 was used as an indication of severe depression. For the Beck Anxiety Inventory, a score over 25 was used as an indicator of severe depression.

Quality of life was measured with the World Health Organization Brief Quality of Life survey (WHO-QoL-Bref) (Skevington, Lotfy, O’Connell, WHOQOL Group, 2004). This is a 26 item self-report measure that includes four domains. The first domain is physical health and asks questions about energy, fatigue, pain, sleep quality, activities of daily living, mobility, and work capacity. The second domain is psychological health and includes questions on bodily image and appearance, negative feelings, positive feelings, self-esteem, spirituality, thinking, memory, and concentration. The third domain queries social relationships including social support and sexual activity. The fourth domain is referred to as environment and includes financial resources, freedom, physical safety, home environment, transportation, opportunities for leisure activities, and aspects of the physical environment such as pollution, noise, traffic, and climate.

The measures of PTSD were the Clinician-Administered PTSD Scale for DSM-5 (CAPS-5 Past Week) (Weathers et al., 2015), the Post-Traumatic Cognitions Inventory (PTCI) (Foa et al., 1999), and the Impact of Events Scales (IES) (Weiss & Marmar, 1996). The CAPS is a 30-item structured interview of symptoms of PTSD. The PTCI is a 33 item self-report measure of negative cognitions about the self, negative cognitions about the world, and self-blame. The IES is a 22 item self-report measure that assesses subjective distress as a result of traumatic events. Both the PTCI and the IES were filled out by the participant. The CAPS is a structured interview administered by an examiner (Weathers et al., 2015). Performance on the CAPS was considered severe if the participant met the criteria for PTSD in the DSM-V. A score of 33 or more on the IES was considered severe.

Survivors of SA were given the HELPS, the ANART, Animal Naming, the Stroop Color-Word Interference Test, the HVLT, the Beck Depression Inventory, the Beck Anxiety Inventory, Impact of Events, and the CAPS. They were also administered the Sexual Assault Exposure Questionnaire (SAEQ) (Rowan et al., 1994).

Procedure

All participants provided written informed consent before any of the testing measures were given. Testing was done in either one or two sessions, depending on participant preference. Total testing time took between one and four hours in a quiet room free from distraction. This was a dedicated testing room on a college campus.

Statistical analysis

Statistical analyses were performed using version 28 of IBM SPSS software (IBM Corp, 2021). Statistical significance was set at p<.05. First demographic (age, education, premorbid intellectual functioning, current occupational status) variables were compared for the three groups (IPV, SA, NV) using one-way analysis of variance (ANOVA) or χ². Second, the three groups (IPV, SA, NV) were compared in terms of social and medical history from the IHHQ. For any significant findings, Fisher’s LSD post-hoc analysis was used to determine which groups differed. Shapiro-Wilk tests were performed and did not show evidence of non-normality (pvalue > .05), therefore parametric tests were used.

In terms of neuropsychological measures, first a series of analyses of covariance (ANCOVA) were performed. To reduce the number of variables, z scores were combined to create three domains of cognitive functioning. The first domain was attention and consisted of the BTA. The second domain was executive functioning and consisted of Animal Naming, COWAT, TMTB, and the Stroop Interference score. The third domain was memory and consisted of the HVLT total score and HVLT delayed recall score. Then a principal components analysis was performed to create a single component for the emotional measures (BAI, BDI, CAPS). A second component was created for history which included family history of domestic violence and family history of alcohol use. The component loadings for each these measures of emotion was .80 or greater. An ANCOVA was run for between group (IPV, SA, NV) differences for each of the three domains (attention, memory, executive functioning) with emotion as a covariate. Assumption evaluations indicated that the normality, homogeneity of variance, linearity and homogeneity of regression slopes assumptions were all satisfactory. Planned comparisons were conducted with the Bryan Paulson Tukey (BPT) test

In order to compare the effect of a blow to the head to NFS, the IPV group was then separated according to whether or not they had reported an incident of NFS. An independent samples t-test was conducted to compare the two groups (IPV with NFS, IPV without NFS) on each of the cognitive measures. Hedges g was used to determine effect size for the t-tests and η² was used to determine effect sizes for the one-way ANOVAs.

Results

Preliminary analyses

There were no significant differences between the groups in terms of age, education, premorbid intelligence, or current occupational status (see Table 1).

Table 1. Demographic information and health history for the three groups.

	IPV Survivors Mean (S.D.)/N (%) N = 50	SA Survivors Mean (S.D.)/N (%) N = 35	NV Mean (S.D.)/N (%) N = 50	F/χ^2
Race/Ethnicity
Black	39%	15%	30%
Hispanic/Latine	42%	23%	35%
White	15%	52%	30%
Other	4%	10%	5%
Medications				0.17
Anti-depressant	3 (6%)	5 (14%)	2 (4%)
Anti-Anxiety	5 (10%)	7 (20%)	2 (4%)
Anti-Psychotic	0	0	0
More than one medication	11 (22%)	6 (17%)	8 (16%)
Alcohol use				43.80**
None	22 (44%)	5 (14%)	9 (18%)
Social	21 (42%)	20 (57%)	29 (58%)
Heavy	2 (4%)	5 (14%)	2 (4%)
Substance Use				2.49
Yes	16 (32%)	18 (51%)	14 (28%)
No	29 (58%)	16 (46 %)	21 (42%)
Family History of Psychiatric Diagnosis	20 (40%)	15 (43%)	7 (14%)	10.92**
Family History of Alcohol Abuse	21 (42%)	10 (29%)	5 (10%)	14.39***
Family History of Eating Disorder	8 (16%)	12 (34%)	1 (2%)	16.26***
Family History of Domestic Violence	22 (44%)	5 (14%)	2 (4%)	25.16***

IPV = intimate partner violence; SA = sexual assault; NV = no violence; S.D. = standard devation.

*p < .05, **p < .01, ***p < .001.

Table 1 also presents the data on the history and family history of the participants. Chi-square analyses revealed significant differences between the groups (IPV, SA, NV) in terms of alcohol use such that IPV survivors were less likely than the other groups to use alcohol. There was no difference between the groups (IPV, SA, NV) in terms of other substance use. There were significant differences between the groups (IPV, SA, NV) in family history such that the IPV survivors were significantly more likely to have a family history of alcohol use and domestic violence.

Detailed information on the specific forms of IPV experienced by the IPV group is presented in Table 2. Fifty-nine percent of the IPV survivors reported having experienced a blow to the head. Thirty-nine percent of the IPV group reported having been shaken and 46% reported having been choked or strangled. Table 2 also presents the number of IPV survivors who had experienced any combination of being hit in the head, strangled, and/or violently shaken.

Table 2. Information on intimate partner violence group.

	Mean	S.D.	Minimum	Maximum	N	% Yes
Duration of intimate partner violence (months)	51.36	59.95	3	240
Months since last event	74.65	135.54	25	504.00
Verbal Abuse					39	78
Sexual Assault					24	48
Threat of Physical Violence					31	62
Physical Violence					31	62
Blow to Head					29	58
With hand					20	40
With head					2	4
With multiple (hand, head, foot, object)					27	54
Times Blow to Head	9.62	21.04	1	100
Whiplash/Shaking					19	38
Times whiplash	8.27	23.30	1	100
Punch/Kick Face					14	28
Choke/Strangle					23	46.0
Times strangle	11.24	39.44	1	50
Blow to head + shaken					20	40
Blow to head + choke/strangle					22	44
Shaken + choke					21	42
Blow to head + shaken + choke					18	36
Hospitalized due to intimate partner violence					9	18
Was the perpetrator prosecuted					16	32
Suicide attempt					9	18

S.D. = standard deviation.

Table 3 presents detailed information on the history of sexual assault by the SA group.

Table 3. Information on sexual assault group.

	Mean	S.D.	Minimum	Maximum	N	% Yes
Duration of sexual assault (months)	3		1	120
Months since last event	35.4		24	60
Exposed to inappropriate comments					32	91
Exposed to someone’s genitals					15	43
Watched while bathig, dressing, or using toilet					5	14
Unwanted touch of genitals					26	74
Someone rubbing genitals against body					5	14
Forced to touch another person					7	20
Forced to have intercourse involving penis/vagina					27	77
Forced to have anal intercourse					2	6
Forced to have oral sex					13	37
Kissed in sexual way against will					12	34
Was the perpetrator prosecuted					2	6
Suicide attempt					4	11

S.D. = standard deviation.

Data from the HELPS brain injury screening tool is presented in Table 4. When asked if they had been hit in the head or had hit their head, 75.6% of IPV survivors said that they had. In terms of overall score on the HELPS, 95% responded yes to three or more items, meeting the criteria for possible brain injury. We then applied the updated HELPS scoring guidelines from the Centers for Disease Control (CDC) (2003). We considered a survivor to be at risk for a TBI if they affirmed the following: (1) A physical injury occurred to the head, neck or face, (2) There was a period of loss of consciousness or altered consciousness after the injury, and (3) Two or more cognitive or physical symptoms were present following the injury. Using these criteria, 60% of this sample screened positive for a potential TBI. We also used these data to apply the ACRM criteria, and this is presented in Table 4. Note that these criteria include specifically, 1. The head being struck, 2. The head striking an object, or 3. An acceleration/deceleration movement, but anoxia is excluded. Detailed brain injury information is presented in Table 4. Table 5 provides self-report information on brain injury related symtpoms.

Table 4. Information on potential brain injury from the HELPS screening tool.

	Percent of IPV survivors who said yes N = 50	Percent of SA survivors who said yes N = 35	Percent of No violence group who said yes N = 50
Hit head or been Hit in the head	75.6%	8%	5%
Emergency room, hospital or doctor because of injury to head	56.1%	3%	1%
Lose consciousness or dazed/confused because of injury to head	62%	5%	5%
Problems in daily life since injury to head	98.8%	2%	0
Sickness since injury to head	65%	2%	2%
Scored 3 or more items on HELPS	95%	0	0
ACRM criteria*	50%	0	0
CDC screening criteria	60%	0	0

ACRM = American College of Rehabilitation Medicine; CDC = Centers for Disease Control; *note anoxia specifically excluded.

IPV = intimate partner violence; SA = sexual assault.

Testing hypotheses

There was a statistically significant difference between the three groups (IPV, SA, NV) on the dependent variable of memory after controlling for the emotion component, F(2, 134) = 31.21, p < .01, partial η² = .491. Planned comparisons indicated significant differences between the IPV and NV groups but not any other groups.

Table 5. Information on brain injury related symptoms reported, including the Rivermead Post-Concussion Questionnaire items.

	IPV survivors (N = 50)		SA survivors (N = 35)		No Violence (N = 50)
	Percent yes	Mean (s.d.)	Percent yes	Mean (s.d.)	Percent yes	Mean (s.d.)
Loss of consciousness	25.6		4.7		2.0
Duration of loss of consciousness (minutes)		12.21 (18.00)		5.1 (6.80)		5.7 (2.78)
Period of confusion after the incident if no loss of consciousness	25.6		0		0
Duration of period of confusion (minutes)		10.43 (13.08)		0		0
Nausea or vomiting after the event	22.0		0		2.1
Light sensitivity after the incident	24.4		0		0
Double vision after the incident	18.2		0		0
Fatigue after incident	24.4		5.7		4.0
Irritability after incident	24.4		0		0
Sensitivity to noise after the incident	22.0		0		0
Sleep disturbance after the incident	24.4		2.1		2.1
Difficulty with balance after the incident	23.2		0		0
Dizziness after the incident	25.6		0		0
Headaches after the incident	35.4		2.8		0
Headache frequency per week		3.76 (2.45)		.5		0
Tinnitus after the incident	9.8		0		0
Difficulty concentrating or remembering after the incident	28.0		0		0

IPV = intimate partner violence; SA = sexual assault; S.D. = standard deviation.

There was also a statistically significant difference between the groups for executive functioning after controlling for emotion, F (2, 134)=35.15, p<.01, partial η²=.24. Planned comparisons revealed significant differences between the IPV and NV groups but no other groups. However, after controlling for emotion there was no significant difference obtained for attention.

When using corrected normative data, χ² indicated significant differences in terms of the number of individuals from each group that met the criteria for impairment only for the HVLT Total Recall. These data are presented in Table 6.

Table 6. Neuropsychological Test performance for intimate partner violence (IPV) group, sexual assault (SA) group, and no violence (NV) group (raw scores, Z scores).

	IPV N = 50			SA N = 35			No V N = 50
	Raw Score Mean (s.d.)	Z Mean (sd)	% (n) impair	Raw Score Mean (s.d.)	Z Mean (sd)	% (n) impair	Raw Score Mean (s.d.)	Z Mean (sd)	% (n) impair	F/η^2	χ^2
ANART	30.11 (10.50)			31.45 (8.75)			31.20 (9.22)			1.03/ .012	0.03
Brief Test of Attention	16.21 (3.19)	−1.24 (0.83)	16 (8)	16. 87 (3.79)	−0.34 (0.12)	14 (5)	17.37 (2.63)	0.12 (0.11)	4 (2)	3.55/ .04	0.06
Animal Naming^a	20.85 (6.50)	−1.02 (0.77)	10 (5)	25.87 (6.91)	0.24 (0.55)	5 (2)	25.03 (5.69)	0.29 (1.12)	0	6.30** .16	0.15
COWAT	24.52 (15.98)	−0.21 (.92)	4 (2)	-------	------	-------	28.41 (15.91)	−0.03 (.56)	2 (1)	1.33/ .05	0.14
TMT B^a (seconds)	70.82 (22.58)	−0.40 (1.45)	12 (6)	58.09 (21.30)	−0.23 (0.55)	3 (1)	55.96 (19.92)	0.43 (0.62)	2 (1)	4.49*/ .07	0.08
Stroop interference^a	42.02 (11.55)	−1.21 (1.89)	18 (9)	46.35 (9.27)	− 0.02 (1.59)	5 (2)	47.87 (9.71)	0.02 (0.78)	2 (1)	4.97*/ .08	0.06
HVLT Total Recall^a	22.97 (5.12)	−1.34 (1.92)	26(13)	30.0 (13.20)	−0.42 (1.49)	17 (6)	25.11 (3.49)	−0.02 (0.19)	6 (3)	6.25*/ .09	0.50
HVLT Delayed Recall^a	7.13 (3.01)	−1.02 (1.79)	30(15)	9.24 (1.28)	−0.55 (1.21)	20 (7)	8.68 (2.12)	−0.18 (0.93)	2 (1)	5.98*/ .02	0.90*
HVLT Retention	75.73 (26.92)	−1.01 (2.98)	4 (2)	92.17 (9.36)	0.25 (4.32)	3 (1)	82.77 (21.92)	1.04 (4.67)	2 (1)	1.25/ .02	0.29

ANOVA F, p, and eta values for raw scores. Z scores derived from published normative data. Impaired defined as >2 standard deviations below the mean and χ² comparing number of individuals impaired in each group.

sd = standard deviation; IPV = intimate partner violence.

^a indicates that Fisher’s LSD demonstrated the significant difference was between IPV and NV groups.

ANART = American National Adult Reading Test; COWAT = Controlled Oral Word Association Test; TMT = Trail Making Test; HVLT = Hopkins Verbal Learning Test; ANOVA = Analysis of variance.

*p < .05; **p < .01.

Table 7 presents the data for the IPV group separated by those who reported NFS and those who did not. Overall, the Student’s t-tests demonstrated significant differences between the raw scores for the groups for the Animal Naming test, the HVLT Total Recall, and the HVLT Delayed Recall. Using corrected normative data and an impairment cut-off of 2 standard deviations below the mean, χ² demonstrated a significant difference in the proportion of individuals who were impaired only for the HVLT Total Recall.

Table 7. Neuropsychological Test performance for intimate partner violence (IPV) group that has been choked and those not choked (raw scores, Z scores).

	IPV with choking N = 26		IPV with no choking N = 24
	Raw Score	% (n) Impaired	Raw Score	% (n) Impaired	t	Hedges g	χ^2
Brief Test of Attention	15.45 (3.58)	19 (5)	17.44 (1.95)	13 (3)	1.12	−0.67	0.42
Animal Naming	19.45 (6.17)	12 (3)	23.11 (6.53)	8 (2)	1.95*	−0.57	0.14
COWAT	27.46 (16.68)	8 (2)	19.94 (14.04)	0	0.86	0.47	0.35
TMT B (seconds)	70.76 (23.45)	19 (5)	71.01 (22.49)	4 (1)	1.02	0.01	0.10
Stroop interference	40.44 (10.42)	23 (6)	44.39 (13.01)	13 (3)	0.95	−0.33	0.95
HVLT Total Recall	21.48 (5.02)	38 (10)	25.39 (4.42)	13 (3)	2.45*	−0.87	4.34**
HVLT Delayed Recall	6.39 (3.07)	38 (10)	8.28 (2.82)	21 (5)	2.29*	0.63	1.85
HVLT Retention	71.14 (31.72)	4 (1)	83.30 (14.06)	4 (1)	1.83	−0.48	0.01

t, p, and cohens’ d values for raw scores. Z scores derived from published normative data. Impaired defined as >2 standard deviations below the mean with χ² to compare number of individuals impaired in each group.

IPV = intimate partner violence; COWAT = Controlled Oral Word Association Test; TMT = Trail Making Test; HVLT = Hopkins Verbal Learning Test.

When comparing proportions of individuals from each group that would be considered to have severe anxiety, depression, and PTSD, χ² indicated significant differences between the groups for depression and for all of the measures of PTSD, as shown in Table 8.

Table 8. Raw scores and % severe for the emotional and quality of life responses for the three groups.

	IPV survivor N = 50		SA Survivor N = 35		No Violence N = 50
	Raw Score	% Severe	Raw Score	% Severe	Raw Score	% Severe	F	η^2	χ^2
Beck Depression Inventory	20.75 (12.35)	23	14.39 (13.66)	15	12.35 (8.55)	0	21.20***	0.79	22.13***
Beck Anxiety Inventory	16.74 (11.36)	24	17.01 (13.45)	22	11.36 (9.32)	12.5	5.09*	0.52	3.87
Impact of Events	38.02 (23.42)	51	22.01 (19.54)	22	18.34 (16.16)	10	16.95***	0.98	33.18***
Post-Traumatic Cognitions Inventory	112.32 (42.25)	32	-----	-----	73.21 (34.09)	5	18.98***	1.02	12.08***
CAPS Total Frequency	5.06 (1.32)	43^b		10	2.22 (0.58)	0	22.11***	1.11	58.43***
WHO Physical (max 35)	21.49 (6.08)		------	-----	22.00 (4.20)		0.18	0.09
WHO Psychological (max 30)	19.46 (6.42)		------	-----	20.51 (3.38)		0.76	1.83
WHO Social (max 15)	9.42 (3.31)		-----	-----	10.60 (2.41)		3.10	0.41
WHO 4 Environment (max 35)	24.25 (7.08)		-----	-----	30.48 (4.28)		20.81***	1.06

F statistic indicates differences between raw scores; χ² indicates differences between percent severe.

^bindicates criteria met for Diagnostic and Statistical Manual-V for post-traumatic stress disorder.

*p < .05; **p < .01; ***p < .001.

LOC was used as a proxy for brain injury for the IPV group. Pearson’s product moment correlations between LOC and cognitive measures indicated that duration of LOC was negatively correlated with HVLT delayed recall (r=-0.70 p<.05) and BTA (-0.72 p<.01) such that longer duration of LOC resulted in lower performance on these measures.

LOC was the only variable that significantly predicted cognitive performance, b = 1.93, t(45) = 6.53, p < .01. LOC also explained a significant proportion of variance in cognitive performance, R² = .12, F(1, 45) = 10.64, p < .01.

Discussion

The current study investigated cognitive performance of a group of women who have experienced IPV compared to a non-IPV comparison group and a group of SA survivors. We looked separately at the effects of blows to the head compared to NFS. A focus of the study was in documenting the occurrence of a potential brain injury in this population as well as adding to current knowledge about the specific cognitive profile demonstrated by this group.

Although there has been increased interest in occult TBI in survivors of IPV in recent years, there are still a small number of studies. This is likely due, in part, to the dual difficulty of low reporting of IPV and low levels of seeking medical care in cases of IPV. Women who experience IPV have low levels of seeking medical help due to fear of their attacker, lack of ability to move freely in the community, fear of leaving their children alone, and stigma associated with IPV (e.g. Sorenson et al., 2014). Just as many providers are not aware of the potential of TBI, so are many of the women who are victims (Rose et al., 2011). In recent reports, only approximately 50% of victims sought medical attention for their injuries (Campbell et al., 2018; Zieman et al., 2017). In fact, the majority of the women in our study (45/50) did not report having knowledge of the possibility of brain injury and only about half (56%) had sought medical care, mostly because of the need for emergency room intervention. This highlights the need for health care providers, social service workers, shelter employees, lawyers, and others to screen for potential brain injury when working with women who have experienced IPV (e.g. Goldin et al., 2016; Haag et al., 2019).

Using the HELPS to screen for brain injury we found that 95% of our IPV sample scored 3 or more items on the HELPS which is comparable to Valera and Kucyi (2017). Using the updated Centers for Disease Control (2003) criteria (see methods for a description), 60% screened as having a possible TBI. This is high compared to 50% reported by Hunnicutt et al. (2019). This may be due to differences in samples. The Hunnicutt et al. (2019) study included anyone who had experienced IPV, including any form of physical, emotional, psychological, verbal, and/or sexual abuse in the context of a relationship with an intimate partner, whereas the current study included primarily women who required shelter placement which may skew it towards more severe physical abuse. In addition, the Hunnicutt et al. (2019) study only surveyed women two years after the abuse so that it is possible that recall of events is not as accurate, or the current study was impacted by recency of the violence and current living situation. None of the participants from the other two groups scored 3 or more on the HELPS. Of course, the HELPS was not designed to diagnose a brain injury but only to serve as an initial screening tool. Thus, positive findings on the HELPS need to be followed up with more comprehensive neuropsychological evaluations.

In addition to the HELPS screening there are several lines of evidence that suggest the possibility of TBI in some members of the IPV group and to suggest that learning and memory may be the most sensitive measures. First, duration of LOC was significantly related to performance on the BTA and the HVLT and was the only significant predictor of cognitive performance in this group. Second, no individuals in the SA group met the criteria for TBI from the HELPS. Third, using raw scores, the group with IPV scored significantly lower than the other two groups, including the SA group with a trauma history, on cognitive measures of executive functioning and memory.

In terms of cognitive functioning, women with a history of IPV demonstrated significantly lower raw score performance on measures of verbal fluency, executive functioning, and memory recall compared to women with a history of SA. This is consistent with the literature on mTBI in general (e.g. Raskin & Mateer, 1999; Raskin et al., 2014) although it is somewhat surprising that there were not lower scores observed on measures of attention. In contrast, LOC did correlate with the measures of attention (BTA) and memory (HVLT delayed recall), suggesting these may be more sensitive measures for determining the occurrence of TBI in IPV. The finding that the HVLT scales were low is consistent with Valera and Berenbaum (2003) who demonstrated impairment on the similar CVLT. It should be noted, however, than clinically there were no differences in the numbers of individuals who demonstrated impairment using a cut-off of more than 2 standard deviations below the published corrected normative scores. Thus, although there may be mild changes in performance these changes do not appear to be widespread severe impairments. Further these findings suggest that the HVLT or a similar list-learning task may be the most sensitive single measure in this population for clinicians wanting to screen for cognitive deficits.

Moreover, the findings of cognitive deficits cannot be explained entirely by the presence of PTSD or negative emotion. When controlling for the measures of emotion (PTSD, depression, and anxiety) significant differences were still found for memory performance on the HVLT and for executive functioning performance as measured by the TMT, verbal fluency, and Stroop word color interference. However, it should be noted that the IPV group did, in fact, demonstrate significantly higher levels of depression, anxiety and PTSD than the non-IPV group. Thus while it is certainly likely that emotional changes contribute to the overall constellation of symptoms (e.g., Kwako et al., 2011), it is important to recognize the unique contribution of cognitive deficits secondary to possible TBI, separate from any existing emotional changes, when creating a comprehensive treatment plan, including interventions for symptoms such as headache and fatigue.

This is consistent with previous research on TBI and PTSD. In previous studies of PTSD (Twamley et al., 2004; Twamley et al., 2009) there were no significant differences found between people with PTSD and those without PTSD on measures of verbal learning, suggesting that lower scores on the HVLT cannot be solely explained by PTSD. In a study looking specifically at survivors of IPV as a PTSD group, the only difference between those with PTSD and healthy participants was on tasks that relied on speed, the Trail Making Test Letter Sequencing and Motor Test, although the performance of the PTSD group was still in the average range (Twamley et al., 2009). They found no differences between the groups on the Stroop or tasks of fluency. Second, in a study that directly compared survivors of childhood SA to those with mTBI, Raskin (1997) found a unique pattern of cognitive deficits in the group that experienced both SA and TBI that differentiated them from groups that experienced only one or the other. Thus, it is possible that in survivors of IPV there will be a similarly unique cognitive profile.

An important limitation of the study is the lack of an appropriate comparison group. The SA group did experience trauma and did not have a history of blows to the head or NFS. However, many of them experienced a single assault and often from someone they did not know well. This is a very different experience from long-term abuse that may go on for years and that involves betrayal by a loved one. Future studies could sample from a broader population of individuals with IPV so that some who did not experience physical violence could be compared to those that did. Unfortunately, we did not have sufficient numbers of those who did not experience physical violence to make that comparison. Additionally, it might be worthwhile to compare women who experienced physical violence after IPV to those who experience mTBI from other causes. Finally, although we did not find an effect of family history of domestic violence or alcohol use, this needs to be investigated more thoroughly. Our finding that the group who had experienced IPV used significantly less alcohol than the other groups should also be replicated and investigated.

When the groups were separated to examine the effects of NFS, those with NFS demonstrated significantly lower performance when compared to the IPV group that did not experience NFS, on measures of verbal fluency, and memory recall when comparing raw scores. This may be because choking leads to an anoxic brain injury or because choking is a marker for more severe abuse in general. The group with NFS, not surprisingly, also had higher rates of emotional complaints which may have contributed to cognitive deficits. This is consistent with and expands the findings of Valera et al. (2022) who found deficits on the CVLT and PTSD measures after NFS. Future studies should look more in depth at questions such as severity of blows to the head, as this study only asked if the woman had been hit in the head. As a result, some individuals may be categorized as TBI based on the HELPS question even if they did not experience a blow severe enough to cause a TBI (i.e. a minor blow to the head in combination with dizziness and confusion from trauma and anxiety). Again, however there was no indication of differences in rates of severe impairment (>2 standard deviations below published normative data) between the groups.

This finding of lower scores in survivors of NFS is consistent with other studies that have found large numbers of women reporting NFS when asked (e.g. Cimino et al., 2019). This speaks to the need for screening measures that include questions about NFS and for greater attention to be paid to these questions. The HELPS includes “near suffocation” under S (sickness). The Conflict Tactics Scale-2 (e.g. Cimino et al., 2019) asks the occurrence and frequency of being “choked” by their partner, although it does not ask if the person lost consciousness as a result. The Danger Assessment Scale has also been modified to include an item asking if the aggressor has ever tried to “choke/strangle” or cut off breathing and whether this lead to dizziness, blacking out or passing out (Messing et al., 2022). Iverson et al. (2017) have modified the VA’s deployment-related TBI screening instrument and this measure includes a question about strangulation or choking. Better would be the creation of a scale to screen for brain injury specifically among IPV survivors. This is, in part, because terminology on NFS is important in screening women survivors. Whereas medically, choke refers to an object within the throat and strangle refers to application of external pressure to the throat, in a focus group of women survivors, they refer to choking as the use of hands externally on their throats and strangle as the use of an object such as a cord or wire (Joshi et al., 2012). Thus, screening measures need to clearly define the question being asked.

Culturally sensitive screening measures for specific populations and in other languages are needed. A further limitation of the current study was the limited scope of the sample. We only evaluated women who were fluent in English and only those connected to shelters. We did not collect adequate data to analyze groups separately by ethnicity or race. Thus, future studies need to look at any differences in rates or types of injury in other populations (e.g. Linton, 2015;  Stockman et al., 2015). In particular, undocumented immigrant women may be at specifically higher risk due to fears of being turned in to immigration authorities.

Another limitation was the heterogeneity of the IPV sample. There were some very recent survivors and some who had experienced IPV many months before. Thus it is very likely that they were mixed in terms of any recovery process with those who are still experiencing acute effects mixed with those who likely had any acute symptoms resolved.

In addition to NFS, attention is needed to understand the potential cumulative effects of repetitive TBIs. All but one of the women in this study would likely be classified as having had a mTBI as only one experienced loss of consciousness for more than 30 min. A limitation of the current study is that we merely queried about number of blows to the head without gathering data on how many of those lead to potential brain injury. However, the average number of blows to the head reported was approximately 10, with a range of 1–100. Given evidence that repeated blows to the head, even at a subconcussive level, can be cumulative (e.g. Gardner & Yaffe, 2015), this is an important area to understand in this population (e.g. Baxter & Hellewell, 2019).

Measuring the repetitive mTBI in this population of women, who are survivors of IPV, is important because of the differences in etiology and mechanism from other types of mTBI (e.g. Raskin et al., 2014). Most previous studies of repetitive TBI have looked at athletes who are in good physical shape, are diagnosed immediately, and receive prompt medical care. Many of these athletes are likely wearing helmets or using other forms of protection against brain injury. Thus, findings from these studies may have limited applicability to women who are IPV survivors. Further, many of the women who had experienced IPV had experienced both a blow to the head and strangulation, or some combination of strangulation, blows, and violent shaking. Therefore, future studies are needed that enroll larger enough sample sizes to allow for comparing the cumulative effects of each combination of violence.

On the other hand, women IPV survivors are more likely to have cognitive deficits lead to immediate negative consequences in their lives. Deficits in memory and executive functioning can impede a survivor’s ability to effectively create and implement a safety plan or to react effectively in a fast-moving situation (e.g. Monahan & O’Leary, 1999). More research is also needed to better understand the reciprocal relationships between these cognitive deficits, symptoms of PTSD, and dysfunction in neural systems (Dossi et al., 2020) which may be cumulative with brain damage from blows to the head or NFS.

Findings on the WHO-QoL-Bref that survivors endorsed a high number of environmental barriers to quality of life highlight the need for future studies to look more in depth at the social inequities that are related to the occurrence of TBI and IPV. For one thing, almost half of the women who experienced IPV in this study reported a history of IPV in their childhood home. Women who are IPV survivors reported significantly higher family history of psychiatric disorders and alcohol abuse. IPV survivors may also have experienced sexual abuse in childhood (St. Ivany & Schminkey, 2016; Suliman et al., 2009; Cimino et al., 2019), which has been shown to lead to potential cognitive impairment itself (Raskin, 1997). In addition, it has been reported that perpetrators of IPV themselves have often experienced a TBI that was not treated (St. Ivany & Schminkey, 2019), with more than 50% of IPV offenders having prior TBIs reported in one systematic review (Farrer et al., 2012).

Acknowledgments

The authors would like to thank all of the women who volunteered their time to be participants and especially the women who are part of our brain injury research advisory board. We would also like to thank Gayna Swart for statistical consultation.

Disclosure statement

We have no known conflict of interest to disclose.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.