Prosodic processing post traumatic brain injury - a systematic review

Over the past 15 years, traumatic brain injuries (TBIs) have been on the rise and have become a public health care issue in North America [1‐10]. The main mechanism of TBI among infants age 0–4, are falls, among adolescents age 10 to 20 are sport related injuries, and among adults are traffic collisions, falls, and assaults [2‐4, 7, 11, 12]. Fall-related head injuries are approximately four times greater among older adults (over 65) compared to adults less than 65 years old [2‐4]. Based on hospitalized and non-hospitalized records, estimates indicate that more males than females sustain TBIs [1‐4, 11]. TBIs have large societal and economic toll [8, 13] and also affect the individual as TBI affects quality of life, including challenges to the individual’s ability to return to work or school and sustain relationships with family, friends, and community [2‐4, 7].

Symptoms associated with TBI involve sensitivity to sounds (increased irritability with loud and/or high pitched sounds) along with various physiological, cognitive, and emotional symptoms including dizziness, headaches, sensitivity to light, mood changes, irritability, diminished focused attention, and slow reaction time [2, 4, 5, 7, 8, 11, 14]. Post-TBI symptoms are difficult to diagnose and prognosis as to whether the symptoms will persist or lead to adverse conditions is difficult [8, 11, 14]. Cognitive, functional, occupational and physical outcomes, as well as social reintegration post-TBI often present a challenge and are issues that have drawn much scientific attention in recent years. For example, adults with a history of TBI report increased current daily cigarette smoking, use of cannabis and non-medically prescribed opioids have significantly higher odds of being currently diagnosed with ADHD and report a greater number of road-related aggression and traffic collisions [12, 15]. Several population studies of adolescents and adults revealed that history of TBI is associated with increased symptoms of depression, anxiety, negative affect, and suicidal ideation [12, 16, 17, 18]. Furthermore, a systematic review by Rogers and Read [19] found that TBI patients are particularly at risk of developing major depression, generalized anxiety disorder and post-traumatic stress disorder years after TBI. Past research has also shown that, not only do injured individuals exhibit impaired employment and productivity post-TBI [20, 21] but they also show impaired executive functions, exhibit higher levels of depression and use maladaptive coping mechanisms such as escape-avoidance mechanisms [21]. Since prosodic perception and decoding of meaning from prosodic information has been shown to be key to well-being, effective social communication and emotional functioning [22], it may be hypothesized that challenges in emotional decoding of communication through sounds following a TBI may disrupt these mechanisms of emotional, adaptive and social functioning.

Speech prosody refers to the melodic aspect of speech. Its function, like that of melody in music, is to convey the emotional content [22, 23] and is a necessary feature in effective social communication [23, 24]. Emotion perception and induction through speech rely on prosody and the emotional code it shares with other prosodic domains (e.g., music and environmental sounds) [22, 23]. To date, no research has explored the diagnosis and assessment of recovery post-TBI using prosodic processing (e.g., music, environmental). Research evidence does exist, however, to suggest that auditory-related functions such prosody and paralinguistic cues (recognition of emotions conveyed through the melodic aspect of speech) are vital to social communication and are susceptible to impairment following TBI [22, 25, 26]. Patients with TBI have been shown to have difficulties not only with general linguistic and paralinguistic abilities, specifically in understanding and producing paralinguistic cues, but also with displaying problems in understanding and producing basic communication tasks [25, 26]. However, the results of these studies have been restricted by small number of participants. Individuals with TBI also very often show a large range of communicative deficits that cannot be adequately explained by linguistic impairment. For example, even though TBI patients may perform normally on standardized aphasia tests, their ability to manage communicative interactions in their daily lives is greatly impaired [14, 25‐27]. This is not surprising, since communicative abilities rely on linguistic (not being able to understand what is implied as in comprehension of sarcastic utterances), extralinguistic (ability to communicate through gestures), paralinguistic (recognizing emotions conveyed through the prosody), context and conversational aspects of communication [24‐26, 28]. For example, TBI patients have trouble ignoring literal meanings of an utterance in order to comprehend sarcasm as well as make indirect inferential requests by giving hints [29, 30]. Angeleri et al. [24] found that TBI patients performed worse than controls on tasks requiring comprehension and production of extralinguistic, paralinguistic, context and conversational communication. Their performance was also impaired in comprehension and production of pragmatic phenomena such as deceit and irony. Impairments in paralinguistic communication may be a causal factor for antisocial behaviour, poor social relationships and aggressive conduct behaviours often reported by patients with TBI [12, 15, 17, 18]. To our knowledge, to date, no systematic review has examined the association between prosodic processing and history of TBI.

This systematic review aims to fill this gap in the literature by examining this relationship between prosodic processing and TBI. The review examines whether prosodic processing impairments are related to history of TBI, TBI severity, brain injury location and recovery. This review also aims to identify whether any co-occurring psychiatric or mental health issues in addition to prosodic impairments post-TBI, in the studies identified, are also reported. Findings from this review will help ascertain if the literature warrants future exploration of the idea that prosodic processing could act as a neuropsychological marker of TBI to help diagnose and index TBI recovery. Implications for the future use of prosodic processing as a neuropsychological marker of TBI through psychoacoustic manipulations of sounds in all auditory domains are discussed.

Following the review of the databases searched, a total of 5212 records were obtained. Based on the inclusion and exclusion criteria, however, only 206 articles were retained for full-text examination as most articles did not report any assessment of prosodic processing; 8 were chosen to be included in this review. A PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flowchart, shown in Fig. 1, was created to illustrate the number of articles found at each stage of data acquisition and the number of articles that were excluded at each stage. All included studies assessed various impairments in prosody processing. Table 1 displays the study characteristics. All studies were published after year 2000, with three conducted in the USA [33‐35], two in the UK [36, 37], two in Australia [38, 39] and one in both, Canada and USA [40]. Three of the studies were prospective cohorts [33, 34, 40] while the others were retrospective. Four studies compared TBI prosodic performance with individuals who had orthopaedic injuries (control) [33, 34, 36, 40], while the others relied on healthy participants as controls. Sample sizes varied between 17 and 71 participants in each condition.

Table 2 displays the participant characteristics of each study. Three of the studies recruited children aged 8 to 17 while the rest recruited adults, usually with a wide age range of up to 70 years old [33, 34, 40]. Time elapsed between injury and assessment for four of the studies were within 5 years [33, 34, 36, 40], while other studies examined patients as far as 25 years post-injury or did not report specific duration. All eight studies reported the sex of the participants and overall had more male than female participants. Years of education were reported by five articles [35‐39] and were matched between control and TBI participants. Severity of TBI was determined by Glasgow Coma Scale (GCS) scores [41], duration of post-traumatic amnesia (PTA), and loss of consciousness (LOC), sometimes also with confirmation of brain lesion from computed tomography (CT) scans. Specific criteria for determining mild, moderate, and severe TBI varied from study to study or were not reported. Mechanism of injury was also heterogeneous, including motor vehicle accidents (MVA), assaults, falls, blow to the head, bike-/sports-/recreation-related accidents, and work-related injuries [33, 36‐40]. Finally, brain localization of injury was reported in five of the eight studies and included various brain areas [34, 36, 38‐40].

The results of this review indicate that the link between language processing and TBI has been clearly understudied. Yet, all eight studies reviewed here found prosodic processing impairments among participants with a history of TBI, compared to controls. TBI participants showed impaired performance and slower reaction time in identifying emotions from prosody, as well as impaired processing of prosodic information that are muffled, non-sense, competing or in conflict (prosody versus semantics). These results are congruent with past studies examining prosodic processing of TBI patients [44‐47]. Cockrell et al. [48] found that 16% of TBI children in their study had central prosodic processing problems. Bergemalm and Borg [49] also evaluated patients with closed-head injuries using peripheral and central prosodic tests as well as questionnaires about hearing ability and quality of life and found that 68% of patients demonstrated abnormalities on one or more audiometric tests. Fourteen of their 25 patients also scored lower on pure-tone audiometry and/or central audiometric tests, and many showed significant progressive deterioration. Taken together, these results illustrate the importance of assessing TBI participants on various aspects of prosodic processing both during the acute phase of the injury and in later follow-ups.

Interestingly, three of the studies reviewed that assessed, in addition to prosodic impairments and other communication difficulties, reading tests, found impairments in some prosodic processing outcomes and not others [36‐38]. Specifically, while some participants were impaired on verbal fluency and National Reading Test for Adults (NART), others performed no different than controls on the Wechsler Test of Adult Reading (WTAR) and the Boston Diagnostic Aphasia test. Impairments in functional communication (which requires individuals to match linguistic acts with paralinguistic elements), however, were much more prominent among individuals with TBI compared with controls, showing TBI-related prosodic processing deficits. Such findings are not surprising. TBI causes a wide range of communication impairments (e.g. aphasia) that cannot be adequately diagnosed by standard language tests [50‐52]. On the other hand, paralinguistic factors such as tone, intonation, rhythm and prosody are crucial in functional communication [53]. Findings showing that TBI patients exhibit prosodic impairments demonstrate that they are indeed struggling with prosodic processing of paralinguistic cues, which although may not be detected by standard language tests can impede effective communication.

Among all eight studies reviewed, prosodic impairments among TBI patients were demonstrated by impairments in identifying emotions and affective meaning communicated through prosody. Although when processing emotional meaning through auditory cues, both linguistic and prosodic information are used, prosody alone becomes extremely important for accuracy in decoding under conditions of ambiguity, sarcasm or irony or when linguistic information is incongruent with vocal affective cues [54, 55]. Findings from the eight studies reviewed demonstrate that TBI participants are not only impaired in recognizing emotions and affect through prosody but that they rely more heavily on linguistic cues alone in tasks requiring accuracy in decoding emotion and affect through auditory cues when the information delivered through prosody and semantics is incongruent. Adults without history of TBI mostly attend to prosodic (instead of linguistic) cues in these situations, a strategy that is particularly helpful in tasks requiring the detection of intentions such as irony and sarcasm [39, 56, 57]. Since studies on prosodic processing among TBI patients are scarce, it is unclear whether TBI patients attend to the linguistic rather than prosodic cues because they have difficulty processing prosody or because attentional impairments make it difficult to recognize the incongruence and lead to a less appropriate strategy for accuracy in decoding. Prosodic impairments were, however, still evident after controlling for cognitive or executive functions. Finally, both studies that used phonological processing tasks as control tasks for auditory processing deficits found that TBI patients presented with impairments in decoding prosody [33, 34].

Some may argue that impaired performance in recognizing emotion through prosody can be attributed to deficits in the brain’s ability to process emotions regardless of medium of delivery, rather than impairment in prosodic processing per se. Given that different sources of linguistic content, prosodic features, facial expression and body movements must be compared and integrated in some manner during interpersonal events, it is not surprising that cues presented in one modality (e.g. auditory) typically interact with cues presented in another modality (e.g. visual) [58]. However, although processing facial expressions (visual modality) and emotional speech prosody (auditory modality) can interact during information processing, they can also occur independently and may have individualized brain processing paths. For example, Adolphs and Tranel [59] found that while the human amygdala was particularly important for recognizing emotional meaning conveyed through facial expressions, but not prosody, the extra-amygdalar structures in the right hemisphere was particularly important for recognizing emotion conveyed through prosody, and not facial expressions.

Findings on effects produced by severity of GCS, age at injury, time-elapsed since injury and SES are mostly heterogeneous, which points to several implications for future research. First, future studies should be more consistent in collecting and reporting these variables, given that, as we observed during the course of this review, these factors may be related with both communicative and prosodic processing outcomes. Secondly, there is a need for longitudinal investigations assessing auditory and prosodic processing at shorter and consistent periods of time post-injury so that we can ascertain with more certainty if the outcomes observed can be linked with the assumed event that caused it and not confounds (repeated head or physical injuries). Just within the eight studies examined, time elapsed since injury ranged from immediately after injury to 25 years post-event, and age at injury, ranged from 8 to 70 years. Criteria for determining severity of TBI relies on differing scores and combinations of GCS, PTA, LOC and CT scans. Neuroplasticity and the brain's ability to re-learn in potentially adaptive circumstances may reduce the amount of prosodic impairments sustained post injury. Hence longitudinal examinations of prosodic impairments over long periods of time post injury would provide important information on prosodic rehabilitation and should be considered in future studies. Furthermore, it may be important to collect and report any prior music education or music training that participants may have received, as past research has shown that music training and exposure to music can improve prosodic processing and decoding of speech prosody, at least among healthy participants [22, 25, 26].

Studies included in our review found an association between temporal lobe damage, of both hemispheres and prosodic impairments. This is consistent with past neuroimaging research studies that have found that both processing of melodic pitch alterations [60] and processing of prosody of words [61] lead to activation of the anterior and posterior regions of the superior temporal sulci. As both rely on processing of prosodic stimuli and melodic contours instead of verbal or cognitive mechanisms, it is important to examine the prosodic processing aspects of prosodic impairments rather than focusing on the cognitive or verbal mechanisms. Future neuroimaging studies for TBI patients exhibiting prosodic impairments are much needed to identify specific neuro-structures and pathways damaged by TBI that could lead to prosodic processing impairments.

Finally, the association between communication deficits potentially caused by prosodic processing impairments experienced by TBI survivors and psychiatric, behavioural, social and employment challenges is not surprising. Past research has found that individuals with a history of TBI have higher odds of being diagnosed for depression, anxiety or both, ADHD, suicide ideation and attempt, hazardous drinking and unemployment or decreased productivity post-TBI [1, 12, 15, 17, 18, 21, 62]. These issues eventually lead to functional impairments and overall reduction in quality of life [63, 64]. Considering that prosodic processing impairments, communication deficits and emotion recognition deficits are also associated with psychosocial problems [53, 65], it raises the importance of identifying prosodic processing issues at early stages of the injury to reduce medical burden and improve outcomes for survivors of TBI.

This review highlights the importance of assessing TBI patients for prosodic processing impairments. TBI patients not only experience prosodic processing impairments that lead to difficulties in effective communication but such impairments often co-occurred with psychiatric issues, behavioural problems and reported unemployment post-TBI. Prosodic processing impairments appear to be correlated with TBI severity, location of brain lesion and time-elapsed since injury, though some of the correlations were inconsistent and warrant further examination.

It is important that future studies collect and report comprehensive details about severity of TBI, location of brain injury and time elapsed since injury, as they could influence the extent of prosodic processing impairments and recovery long term. If future studies could elucidate correlations between prosodic processing and TBI severity, prosodic processing screening could potentially be a valuable neuropsychological marker of TBI diagnosis and recovery. Prosodic processing is crucial for effective functional communication, and impairments in this domain could be a contributing factor to the many psychological problems and poor social outcomes observed in TBI patients. Treating prosodic processing issues in a timely fashion could result in significantly better functional outcome and help in costs reduction involved in the care for this population.