Hugo Aben

Main findings
In this thesis, we described a case report of a patient that had impaired facial emotion recognition after ischemic stroke located in the left insula (chapter 6). This seemingly contrasts with earlier literature that states that the right 61 hemisphere is responsible for the process of emotion recognition. An analysis of the structural brain network of this case showed that the stroke may have caused a more widespread disruption of his brain network, which may explain why he had more difficulties recognizing facial emotions. Based on a brief literature review, we conclude in chapter 6 that it is in fact a bihemispheral network of neural structures that is involved in correctly recognizing emotions. It may not come as a surprise that when an ischemic stroke causes dysfunction anywhere in this network of structures, emotion recognition may get impaired. In chapter 7 we showed that emotion recognition is in fact impaired in 1 out of 3 patients after ischemic stroke. These patients were actually rather similar to the patients that did not have impaired emotion recognition. This means that in contrast with what was previously thought, impairment in recognizing emotions occurs regardless of the size of stroke and lateralization of stroke. Again, it may be that considering location of ischemic stroke within the specific subnetwork of emotion recognition may further our understanding of occurrence of impaired emotion recognition.

Recently, it was shown that impairments in social cognition remain present in the 62 long term after ischemic stroke. This suggests that patients may show only little recovery in their ability to recognize emotions after stroke. As discussed earlier, impairment in each cognitive domain may have a different and unique impact on functioning in daily life. It may be helpful to also consider emotion recognition recovery in future studies of cognitive recovery after stroke. We conclude that because of the high occurrence of impaired emotion recognition after stroke, it is important for clinicians to evaluate social cognition in daily practice, since it can have negative consequences on social participation and quality of life. 63-66 We propose that routinely asking patients and their caregivers about post-stroke changes in social behavior or empathy can help selecting patients that need further neuropsychological examination of their social cognitive capacities. Moreover, psychoeducation can be provided, as it is an important first step in cognitive 55 rehabilitation. It is important that patients with deficits in social cognition are recognized, because there is evidence in patients with traumatic brain injury that a strategy training for impairments in social cognition can have long-lasting 67 beneficial effects. The effect of this training program is going to be studied in 8 patients with ischemic stroke.

Clinical implications
In the work described in this thesis, we have shown that measures of global brain connectivity look promising, but at this moment do not add prognostic value in predicting cognitive recovery after stroke in a way that can be applied in current clinical practice. However, when combining information on network topology with other modalities (i.e. multimodal measures), this may help better predicting a patient’s chances of cognitive recovery. Before a multimodal measure such as the lesion impact score (chapter 3), can be implemented in daily practice, more research is needed. In the next few years, several hurdles have to be taken to get brain networks in daily practice. First, reliable and strong prognostic models need to be developed. This means that models do not only need to explain a large amount of variance, but also that findings are reproducible across and within different cohorts. Next, it is important to invest resources in automation of most of the post-processing of MRI scans of patients. This would also enable a different output of brain imaging results, where not only a qualitative description of the radiologist, but also quantitative results of, for example, a patient’s brain volume, infarct volume and network properties could be reported in relation to reference values. There is already a substantial amount of published data on the application 68 of this so-called ‘radiomics’ (i.e. quantitative imaging) in cancer research. When it is possible to introduce quantitative imaging results into daily practice, this can be the next step to more individualized tailored care.

The research that was described in chapters 6 and 7 can have implications for daily practice directly. We found that impairments in emotion recognition occur frequently after ischemic stroke. As we know from earlier research, patients with ischemic stroke are not unique in having impairments in emotion recognition: it 60 can be seen in all kinds of neurological and psychiatric diseases. This implies that clinicians should more actively ask their patients and their caregivers about post-stroke changes in social behavior or empathy. Psychoeducation should be provided to patients with signs of impairment in their social cognitive capacities, and a referral to a neuropsychologist should be considered. The best way to achieve this, is to increase awareness by educating all involved caregivers in stroke care on the cognitive and emotion consequences following stroke, such as neurologists, rehabilitation physicians, neuropsychologists, residents neurology, physician assistants, physical and occupational therapists and (specialized) nurses. In the following years, it seems to be important that the patient and their caregiver take as much control over the patient’s complaints as possible. This would mean that it is possible for a patient to get easy access to information and if needed he would, for example, receive screening questions regarding cognition and social cognition through a patient portal. If the patient screens positive, then this would lead to an invitation to the outpatient clinic where the complaint can be studied further.

In the work described in chapter 5, we showed that 5 weeks after stroke patients often do not have a corresponding lesion on MRI. Although a negative MRI should prompt the question whether the patient did suffer from ischemic stroke, we showed that a large proportion of these patients still had a likely clinical diagnosis of ischemic stroke. Patients with a negative MRI were rather similar to patients that had a corresponding lesion on MRI. For MRI negative patients it is just as important to have information on prognosis as for MRI positive patients. When looking at prognosis, patients with MRI negative stroke do not differ much from patients that had a nonspecific transient ischemic attack (TIA), which has been coined as a 69 transient neurological attack (TNA). The similarity between MRI positive stroke as compared to MRI negative stroke and TIA compared to TNA is actually rather striking. Patients with TNA also have a higher risk of recurrent stroke, similar to MRI negative stroke patients. 70-72 In patients with TIA and TNA, diagnosis is based on the judgement of the clinician, as there is no gold standard diagnostic test. If an urgent MRI is made, abnormalities can be seen in approximately 20-30% of patients in both groups. 69,72 Understandably, there is an ongoing discussion as to how to treat patients with TNA, should they be treated likewise to patients 72 with TIA? In my opinion, this discussion could be extended to patients with MRI negative stroke. In absence of a better explanation for the symptoms of a patient with MRI negative stroke, the patient should be treated the same way as patients with MRI positive stroke would be treated. This indicates that clinicians should be aware that a negative MRI is not unusual in patients that suffered from an ischemic stroke, especially in patients with less severe strokes, and less specific symptoms. This finding can have direct consequences for daily practice, as physicians tend to use the MRI to rule-in or rule-out the diagnosis of ischemic stroke. This can lead to patients with MRI negative stroke being undertreated and this would leave them at higher risk of a recurrent stroke. Again, it seems important to increase awareness and to educate physicians on the existence of an MRI negative stroke.

Conclusions
Cognitive impairment occurs often after ischemic stroke. Establishing which patient has the potential to recover, could help in adjusting rehabilitation programs to the patient’s needs and capabilities. We showed that DWI based measures of brain connectivity are promising in determining a patient’s potential to recover. However, further research is needed to assess how these measures can add value 8 in daily practice.

We also showed that emotion recognition is impaired in one out of three patients, while it is often not recognized by the clinician or the patient. Clinicians should be aware that these impairments occur often and should routinely inform patients about changes in social behavior or empathy. If there are signs of changes in a patient’s social behavior or empathy, patients or their caregivers should be provided with psychoeducation and referral to a neuropsychologist should be considered.