GABAa Receptor Mediated Signalling in the Brain: Inhibition, Shunting and Excitation

Within central nervous system, the simple division of chemical synaptic transmission to depolarizing excitation mediated by glutamate and hyperpolarizing inhibition mediated by γ-amino butyric acid (GABA), is evidently an oversimplification. The GABAa receptor (GABAaR) mediated responses can be of opposite sign within a single resting cell, due to the compartmentalized distribution of cation chloride cotransporters (CCCs). The K+/Cl- cotransporter 2 (KCC2), member of the CCC family, promotes K+ fuelled Cl- extrusion and sets the reversal potential of GABA evoked anion currents typically slightly below the resting membrane potential. The interesting ionic plasticity property of GABAergic signalling emerges from the short-term and long-term alterations in the intraneuronal concentrations of GABAaR permeable anions (Cl- and HCO3-). The short-term effects arise rapidly (in the time scale of hundreds of milliseconds) and are due to the GABAaR activation dependent shifts in anion gradients, whereas the changes in expression, distribution and kinetic regulation of CCCs are underlying the long-term effects, which may take minutes or even hours to develop. In this Thesis, the differences in the reversal potential of GABAaR mediated responses between dopaminergic and GABAergic cell types, located in the substantia nigra, were shown to be attributable to the differences in the chloride extrusion mechanisms. The stronger inhibitory effect of GABA on GABAergic neurons was due to the cell type specific expression of KCC2 whereas the KCC2 was absent from dopaminergic neurons, leading to a less prominent inhibition brought by GABAaR activation. The levels of KCC2 protein exhibited activity dependent alterations in hippocampal pyramidal neurons. Intense neuronal activity, leading to a massive release of brain derived neurotrophic factor (BDNF) in vivo, or applications of tyrosine receptor kinase B (TrkB) agonists BDNF or neurotrophin-4 in vitro, were shown to down-regulate KCC2 protein levels which led to a reduction in the efficacy of Cl- extrusion. The GABAergic transmission is interestingly involved in an increase of extracellular K+ concentration. A substantial increase in interstitial K+ tends to depolarize the cell membrane. The effects that varying ion gradients had on the generation of biphasic GABAaR mediated responses were addressed, with particular emphasis on the novel idea that the K+/Cl- extrusion via KCC2 is accelerated in response to a rapid accumulation of intracellular Cl-. The KCC2 inhibitor furosemide produced a large reduction in the GABAaR dependent extracellular K+ transients. Thus, paradoxically, both the inefficient KCC2 activity (via increased intracellular Cl-) and efficient KCC2 activity (via increased extracellular K+) may promote excitation.

Neuromagnetic studies on cortical somatosensory functions in infants and children : Normal development and effect of early brain lesions

Until recently, objective investigation of the functional development of the human brain in vivo was challenged by the lack of noninvasive research methods. Consequently, fairly little is known about cortical processing of sensory information even in healthy infants and children. Furthermore, mechanisms by which early brain insults affect brain development and function are poorly understood. In this thesis, we used magnetoencephalography (MEG) to investigate development of cortical somatosensory functions in healthy infants, very premature infants at risk for neurological disorders, and adolescents with hemiplegic cerebral palsy (CP). In newborns, stimulation of the hand activated both the contralateral primary (SIc) and secondary somatosensory cortices (SIIc). The activation patterns differed from those of adults, however. Some of the earliest SIc responses, constantly present in adults, were completely lacking in newborns and the effect of sleep stage on SIIc responses differed. These discrepancies between newborns and adults reflect the still developmental stage of the newborns’ somatosensory system. Its further maturation was demonstrated by a systematic transformation of the SIc response pattern with age. The main early adult­like components were present by age two. In very preterm infants, at term age, the SIc and SIIc were activated at similar latencies as in healthy fullterm newborns, but the SIc activity was weaker in the preterm group. The SIIc response was absent in four out of the six infants with brain lesions of the underlying hemisphere. Determining the prognostic value of this finding remains a subject for future studies, however. In the CP adolescents with pure subcortical lesions, contrasting their unilateral symptoms, the SIc responses of both hemispheres differed from those of controls: For example the distance between SIc representation areas for digits II and V was shorter bilaterally. In four of the five CP patients with cortico­subcortical brain lesions, no normal early SIc responses were evoked by stimulation of the palsied hand. The varying differences in neuronal functions, underlying the common clinical symptoms, call for investigation of more precisely designed rehabilitation strategies resting on knowledge about individual functional alterations in the sensorimotor networks.

Effects of opioids on ventilation and hemodynamics

Opioids are most commonly used for treatment of severe pain. However, the fear of respiratory depression has restricted the use of opioids. Depending on the monitoring system used, different modes of opioid respiratory effects have been noted in previous studies. All opioids also cause alterations in hemodynamics at least to some extent. The main goal of this series of investigations was to elucidate the native ventilatory and hemodynamic effects of different opioids. Studies I-IV each involved 8 healthy male volunteers. Study V involved 13 patients with lower or upper extremity traumas. The opioids studied were morphine, oxycodone, pethidine, fentanyl, alfentanil, tramadol and ketamine. The respiratory parameters used in this study were breathing pattern measured with respiratory inductive plethysmography, gas exchange measured with indirect calorimetry, blood gas analysis and pulse oximetry. Hemodynamics was measured with arterial blood pressure, heart rate and oxygen consumption. Plasma catecholamine and histamine concentrations were also determined. All opioids studied caused an alteration in respiratory function. Respiratory rate, alveolar ventilation and minute ventilation decreased, while tidal volume increased in most situations. Breathing pattern was also significantly affected after opioid administration. The respiratory depression caused by oxycodone was deeper than the one caused by same dose of morphine. An equianalgesic dose of tramadol caused markedly smaller respiratory depression compared to pethidine. The potency ratio for respiratory depression of fentanyl and alfentanil is similar to analgesic potency ratio studied elsewhere. Racemic ketamine attenuated the respiratory depression caused by fentanyl, if measured with minute ventilation. However, this effect was counteracted by increased oxygen consumption. Supplemental oxygen did not offer any benefits, nor did it cause any atelectasis when given to opioid treated trauma patients. Morphine caused a transient hemodynamic stimulation, which was accompanied by an increase in oxygen consumption. Oxycodone, alfentanil, fentanyl, tramadol and pethidine infusions had minimal effects on hemodynamics. Plasma catecholamine concentrations were increased after high dose opioid administration. Plasma histamine concentrations were not elevated after morphine nor oxycodone administration. Respiratory depression is a side effect noted with all opioids. The profile of this phenomenon is quite similar with different opioid-receptor agonists. The hemodynamic effects of opioids may vary depending on the opioid used, morphine causing a slight hemodynamic stimulation. However, all opioids studied could be considered hemodynamically stable.

The brain serotonin transporter binding in young adults; methodological considerations and association with Bulimia Nervosa and acquired obesity

The neurotransmitter serotonin (5-HT) modulates many functions important for life, e.g., appetite and body temperature, and controls development of the neural system. Disturbed 5-HT function has been implicated in mood, anxiety and eating disorders. The serotonin transporter (SERT) controls the amount of effective 5-HT by removing it from the extracellular space. Radionuclide imaging methods single photon emission tomography (SPET) and positron emission tomography (PET) enable studies on the brain SERTs. This thesis concentrated on both methodological and clinical aspects of the brain SERT imaging using SPET. The first study compared the repeatability of automated and manual methods for definition of volumes of interest (VOIs) in SERT images. The second study investigated within-subject seasonal variation of SERT binding in healthy young adults in two brain regions, the midbrain and thalamus. The third study investigated the association of the midbrain and thalamic SERT binding with Bulimia Nervosa (BN) in female twins. The fourth study investigated the association of the midbrain and hypothalamic/thalamic SERT binding and body mass index (BMI) in monozygotic (MZ) twin pairs. Two radioligands for SERT imaging were used: [123I]ADAM (studies I-III) and [123I]nor-beta-CIT (study IV). Study subjects included young adult MZ and dizygotic (DZ) twins screened from the FinnTwin16 twin cohort (studies I-IV) and healthy young adult men recruited for study II. The first study validated the use of an automated brain template in the analyses of [123I]ADAM images and proved automated VOI definition more reproducible than manual VOI definition. The second study found no systematic within-subject variation in SERT binding between scans done in summer and winter in either of the investigated brain regions. The third study found similar SERT binding between BN women (including purging and non-purging probands), their unaffected female co-twins and other healthy women in both brain regions; in post hoc analyses, a subgroup of purging BN women had significantly higher SERT binding in the midbrain as compared to all healthy women. In the fourth study, MZ twin pairs were divided into twins with higher BMI and co-twins with lower BMI; twins with higher BMI were found to have higher SERT binding in the hypothalamus/thalamus than their leaner co-twins. Our results allow the following conclusions: 1) No systematic seasonal variation exists in the midbrain and thalamus between SERT binding in summer and winter. 2) In a population-based sample, BN does not associate with altered SERT status, but alterations are possible in purging BN women. 3) The higher SERT binding in MZ twins with higher BMIs as compared to their leaner co-twins suggests non-genetic association between acquired obesity and the brain 5-HT system, which may have implications on feeding behavior and satiety.

Therapeutic hypothermia after cardiac arrest : Studies on neurological and cardiological outcome and prediction of outcome in hypothermia-treated patients resuscitated from out-of-hospital cardiac arrest

The outcome of the successfully resuscitated patient is mainly determined by the extent of hypoxic-ischemic cerebral injury, and hypothermia has multiple mechanisms of action in mitigating such injury. The present study was undertaken from 1997 to 2001 in Helsinki as a part of the European multicenter study Hypothermia after cardiac arrest (HACA) to test the neuroprotective effect of therapeutic hypothermia in patients resuscitated from out-of-hospital ventricular fibrillation (VF) cardiac arrest (CA). The aim of this substudy was to examine the neurological and cardiological outcome of these patients, and especially to study and develop methods for prediction of outcome in the hypothermia-treated patients. A total of 275 patients were randomized to the HACA trial in Europe. In Helsinki, 70 patients were enrolled in the study according to the inclusion criteria. Those randomized to hypothermia were actively cooled externally to a core temperature 33 ± 1ºC for 24 hours with a cooling device. Serum markers of ischemic neuronal injury, NSE and S-100B, were sampled at 24, 36, and 48 hours after CA. Somatosensory and brain stem auditory evoked potentials (SEPs and BAEPs) were recorded 24 to 28 hours after CA; 24-hour ambulatory electrocardiography recordings were performed three times during the first two weeks and arrhythmias and heart rate variability (HRV) were analyzed from the tapes. The clinical outcome was assessed 3 and 6 months after CA. Neuropsychological examinations were performed on the conscious survivors 3 months after the CA. Quantitative electroencephalography (Q-EEG) and auditory P300 event-related potentials were studied at the same time-point. Therapeutic hypothermia of 33ºC for 24 hours led to an increased chance of good neurological outcome and survival after out-of-hospital VF CA. In the HACA study, 55% of hypothermia-treated patients and 39% of normothermia-treated patients reached a good neurological outcome (p=0.009) at 6 months after CA. Use of therapeutic hypothermia was not associated with any increase in clinically significant arrhythmias. The levels of serum NSE, but not the levels of S-100B, were lower in hypothermia- than in normothermia-treated patients. A decrease in NSE values between 24 and 48 hours was associated with good outcome at 6 months after CA. Decreasing levels of serum NSE but not of S-100B over time may indicate selective attenuation of delayed neuronal death by therapeutic hypothermia, and the time-course of serum NSE between 24 and 48 hours after CA may help in clinical decision-making. In SEP recordings bilaterally absent N20 responses predicted permanent coma with a specificity of 100% in both treatment arms. Recording of BAEPs provided no additional benefit in outcome prediction. Preserved 24- to 48-hour HRV may be a predictor of favorable outcome in CA patients treated with hypothermia. At 3 months after CA, no differences appeared in any cognitive functions between the two groups: 67% of patients in the hypothermia and 44% patients in the normothermia group were cognitively intact or had only very mild impairment. No significant differences emerged in any of the Q-EEG parameters between the two groups. The amplitude of P300 potential was significantly higher in the hypothermia-treated group. These results give further support to the use of therapeutic hypothermia in patients with sudden out-of-hospital CA.

The role of mast cells in ischemic and hemorrhagic brain injury

Stroke, ischemic or hemorrhagic, belongs among the foremost causes of death and disability worldwide. Massive brain swelling is the leading cause of death in large hemispheric strokes and is only modestly alleviated by available treatment. Thrombolysis with tissue plasminogen activator (TPA) is the only approved therapy in acute ischemic stroke, but fear of TPA-mediated hemorrhage is often a reason for withholding this otherwise beneficial treatment. In addition, recanalization of the occluded artery (spontaneously or with thrombolysis) may cause reperfusion injury by promoting brain edema, hemorrhage, and inflammatory cell infiltration. A dominant event underlying these phenomena seems to be disruption of the blood-brain barrier (BBB). In contrast to ischemic stroke, no widely approved clinical therapy exists for intracerebral hemorrhage (ICH), which is associated with poor outcome mainly due to the mass effect of enlarging hematoma and associated brain swelling. Mast cells (MCs) are perivascularly located resident inflammatory cells which contain potent vasoactive, proteolytic, and fibrinolytic substances in their cytoplasmic granules. Experiments from our laboratory showed MC density and their state of granulation to be altered early following focal transient cerebral ischemia, and degranulating MCs were associated with perivascular edema and hemorrhage. (I) Pharmacological MC stabilization led to significantly reduced ischemic brain swelling (40%) and BBB leakage (50%), whereas pharmacological MC degranulation raised these by 90% and 50%, respectively. Pharmacological MC stabilization also revealed a 40% reduction in neutrophil infiltration. Moreover, genetic MC deficiency was associated with an almost 60% reduction in brain swelling, 50% reduction in BBB leakage, and 50% less neutrophil infiltration, compared with controls. (II) TPA induced MC degranulation in vitro. In vivo experiments with post-ischemic TPA administration demonstrated 70- to 100-fold increases in hemorrhage formation (HF) compared with controls HF. HF was significantly reduced by pharmacological MC stabilization at 3 (95%), 6 (75%), and 24 hours (95%) of follow-up. Genetic MC deficiency again supported the role of MCs, leading to 90% reduction in HF at 6 and 24 hours. Pharmacological MC stabilization and genetic MC deficiency were also associated with significant reduction in brain swelling and in neutrophil infiltration. Importantly, these effects translated into a significantly better neurological outcome and lower mortality after 24 hours. (III) Finally, in ICH experiments, pharmacological MC stabilization resulted in significantly less brain swelling, diminished growth in hematoma volume, better neurological scores, and decreased mortality. Pharmacological MC degranulation produced the opposite effects. Genetic MC deficiency revealed a beneficial effect similar to that found with pharmacological MC stabilization. In sum, the role of MCs in these clinically relevant scenarios is supported by a series of experiments performed both in vitro and in vivo. That not only genetic MC deficiency but also drugs targeting MCs could modulate these parameters (translated into better outcome and decreased mortality), suggests a potential therapeutic approach in a number of highly prevalent cerebral insults in which extensive tissue injury is followed by dangerous brain swelling and inflammatory cell infiltration. Furthermore, these experiments could hint at a novel therapy to improve the safety of thrombolytics, and a potential cellular target for those seeking novel forms of treatment for ICH.

Carotid Stenosis, Endarterectomy, and the Brain : Brain microcirculation, diffusion, and cognitive function before and after carotid endarterectomy in patients with a high-grade carotid stenosis

Carotid atherosclerotic disease is a major cause of stroke, but it may remain clinically asymptomatic. The factors that turn the asymptomatic plaque into a symptomatic one are not fully understood, neither are the subtle effects that a high-grade carotid stenosis may have on the brain. The purpose of this study was to evaluate brain microcirculation, diffusion, and cognitive performance in patients with a high-grade stenosis in carotid artery, clinically either symptomatic or asymptomatic, undergoing carotid endarterectomy (CEA). We wanted to find out whether the stenoses are associated with diffusion or perfusion abnormalities of the brain or variation in the cognitive functioning of the patients, and to what extent the potential findings are affected by CEA, and compare the clinically symptomatic and asymptomatic subjects as well as strictly healthy controls. Coagulation and fibrinolytic parameters were compared with the rate microembolic signals (MES) in transcranial Doppler (TCD) and the macroscopic appearance of stenosing plaques in surgery. Patients (n=92) underwent CEA within the study. Blood samples pertaining to coagulation and fibrinolysis were collected before CEA, and the subjects underwent repeated TCD monitoring for MES. A subpopulation (n= 46) underwent MR imaging and repeated neuropsychological examination (preoperative, as well 4 and 100 days after CEA). In MRI, the average apparent diffusion coefficients were higher in the ipsilateral white matter (WM), and altough the interhemispheric difference was abolished by CEA, the levels remained higher than in controls. Symptomatic stenoses were associated with more sluggish perfusion especially in WM, and lower pulsatility of flow in TCD. All patients had poorer cognitive performance than healthy controls. Cognitive functions improved as expected by learning effect despite transient postoperative worsening in a few subjects. Improvement was greater in patients with deepest hypoperfusion, primarily in executive functions. Symptomatic stenoses were associated with higher hematocrit and tissue plasminogen activator antigen levels, as well as higher rate of MES and ulcerated plaques, and better postoperative improvement of vasoreactivity and pulsatility. In light of the findings, carotid stenosis is associated with differences in brain diffusion, perfusion, and cognition. The effect on diffusion in the ipsilateral WM, partially reversible by CEA, may be associated with WM degeneration. Asymptomatic and symptomatic subpopulations differ from each other in terms of hemodynamic adaptation and in their vascular physiological response to removal of stenosis. Although CEA may be associated with a transient cognitive decline, a true improvement of cognitive performance by CEA is possible in patients with the most pronounced perfusion deficits. Mediators of fibrinolysis and unfavourable hemorheology may contribute to the development of a symptomatic disease in patients with a high-grade stenosis.

Effects of Metabolic Risk Factors and Type 1 Diabetes on Brain Glucose and Brain Metabolites

The metabolic syndrome and type 1 diabetes are associated with brain alterations such as cognitive decline brain infarctions, atrophy, and white matter lesions. Despite the importance of these alterations, their pathomechanism is still poorly understood. This study was conducted to investigate brain glucose and metabolites in healthy individuals with an increased cardiovascular risk and in patients with type 1 diabetes in order to discover more information on the nature of the known brain alterations. We studied 43 20- to 45-year-old men. Study I compared two groups of non-diabetic men, one with an accumulation of cardiovascular risk factors and another without. Studies II to IV compared men with type 1 diabetes (duration of diabetes 6.7 ± 5.2 years, no microvascular complications) with non-diabetic men. Brain glucose, N-acetylaspartate (NAA), total creatine (tCr), choline, and myo-inositol (mI) were quantified with proton magnetic resonance spectroscopy in three cerebral regions: frontal cortex, frontal white matter, thalamus, and in cerebellar white matter. Data collection was performed for all participants during fasting glycemia and in a subgroup (Studies III and IV), also during a hyperglycemic clamp that increased plasma glucose concentration by 12 mmol/l. In non-diabetic men, the brain glucose concentration correlated linearly with plasma glucose concentration. The cardiovascular risk group (Study I) had a 13% higher plasma glucose concentration than the control group, but no difference in thalamic glucose content. The risk group thus had lower thalamic glucose content than expected. They also had 17% increased tCr (marker of oxidative metabolism). In the control group, tCr correlated with thalamic glucose content, but in the risk group, tCr correlated instead with fasting plasma glucose and 2-h plasma glucose concentration in the oral glucose tolerance test. Risk factors of the metabolic syndrome, most importantly insulin resistance, may thus influence brain metabolism. During fasting glycemia (Study II), regional variation in the cerebral glucose levels appeared in the non-diabetic subjects but not in those with diabetes. In diabetic patients, excess glucose had accumulated predominantly in the white matter where the metabolite alterations were also the most pronounced. Compared to the controls values, the white matter NAA (marker of neuronal metabolism) was 6% lower and mI (glia cell marker) 20% higher. Hyperglycemia is therefore a potent risk factor for diabetic brain disease and the metabolic brain alterations may appear even before any peripheral microvascular complications are detectable. During acute hyperglycemia (Study III), the increase in cerebral glucose content in the patients with type 1 diabetes was, dependent on brain region, between 1.1 and 2.0 mmol/l. An every-day hyperglycemic episode in a diabetic patient may therefore as much as double brain glucose concentration. While chronic hyperglycemia had led to accumulation of glucose in the white matter, acute hyperglycemia burdened predominantly the gray matter. Acute hyperglycemia also revealed that chronic fluctuation in blood glucose may be associated with alterations in glucose uptake or in metabolism in the thalamus. The cerebellar white matter appeared very differently from the cerebral (Study IV). In the non-diabetic men it contained twice as much glucose as the cerebrum. Diabetes had altered neither its glucose content nor the brain metabolites. The cerebellum seems therefore more resistant to the effects of hyperglycemia than is the cerebrum.

Brain and Spinal Cavernomas : Helsinki Experience

Cavernomas are rare neurovascular lesions, encountered in up to 10% of patients harboring vascular abnormalities of the CNS. Cavernomas consist of dilated thin-walled sinusoids or caverns covered by a single layer of endothelium. Due to advancements in neuroradiology, the number of cavernoma patients coming to be evaluated in neurosurgical practice is increasing. In the present work, we summarized our results on the treatment of cavernomas. Particular attention was paid to uncommon locations or insufficiently investigated cavernomas, including 1. Intraventricular cavernomas; 2. Multiple cavernomas; 3. Spinal cavernomas; and 4. Temporal lobe cavernomas. After analyzing the patient series with these lesions, we concluded that: 1. IVCs are characterized by a high tendency to cause repetitive hemorrhages in a short period of time after the first event. In most patients, hemorrhages were not life-threatening. Surgery is indicated when re-bleedings are frequent and the mass-effect causes progressive neurological deterioration. Modern microsurgical techniques allow safe removal of the IVC, but surgery on fourth ventricle cavernomas carries increased risk of postoperative cranial nerve deficits. 2. In MC cases, when the cavernoma bleeds or generates drug-resistant epilepsy, microsurgical removal of the symptomatic lesion is beneficial to patients. In our series, surgical removal of the most active cavernoma usually the biggest lesion with signs of recent hemorrhage - was safe and prevented further bleedings. Epilepsy outcome showed the effectiveness of active treatment of MCs. However, due to the remaining cavernomas, epileptogenic activity can persist postoperatively, frequently necessitating long-term use of antiepileptic drugs. 3. Spinal cavernomas can cause severe neurological deterioration due to low tolerance of the spinal cord to mass-effect with progressive myelopathy. When aggravated by extralesional massive hemorrhage, neurological decline is usually acute and requires immediate treatment. Microsurgical removal of a cavernoma is effective and safe, improving neurological deficits. Sensorimotor deficits and pain improved postoperatively at a high rate, whereas bladder dysfunction remained essentially unchanged, causing social discomfort to patients. 4. Microsurgical removal of temporal lobe cavernomas is beneficial for patents suffering from drug-resistant epilepsy. In our series, 69% of patients with this condition became seizure-free postoperatively. Duration of epilepsy did not correlate with seizure prognosis. The most frequent disabling symptom at follow-up was memory disorder, considered to be the result of a complex interplay between chronic epilepsy and possible damage to the temporal lobe during surgery.

Language learning in infancy

Although immensely complex, speech is also a very efficient means of communication between humans. Understanding how we acquire the skills necessary for perceiving and producing speech remains an intriguing goal for research. However, while learning is likely to begin as soon as we start hearing speech, the tools for studying the language acquisition strategies in the earliest stages of development remain scarce. One prospective strategy is statistical learning. In order to investigate its role in language development, we designed a new research method. The method was tested in adults using magnetoencephalography (MEG) as a measure of cortical activity. Neonatal brain activity was measured with electroencephalography (EEG). Additionally, we developed a method for assessing the integration of seen and heard syllables in the developing brain as well as a method for assessing the role of visual speech when learning phoneme categories. The MEG study showed that adults learn statistical properties of speech during passive listening of syllables. The amplitude of the N400m component of the event-related magnetic fields (ERFs) reflected the location of syllables within pseudowords. The amplitude was also enhanced for syllables in a statistically unexpected position. The results suggest a role for the N400m component in statistical learning studies in adults. Using the same research design with sleeping newborn infants, the auditory event-related potentials (ERPs) measured with EEG reflected the location of syllables within pseudowords. The results were successfully replicated in another group of infants. The results show that even newborn infants have a powerful mechanism for automatic extraction of statistical characteristics from speech. We also found that 5-month-old infants integrate some auditory and visual syllables into a fused percept, whereas other syllable combinations are not fully integrated. Auditory syllables were paired with visual syllables possessing a different phonetic identity, and the ERPs for these artificial syllable combinations were compared with the ERPs for normal syllables. For congruent auditory-visual syllable combinations, the ERPs did not differ from those for normal syllables. However, for incongruent auditory-visual syllable combinations, we observed a mismatch response in the ERPs. The results show an early ability to perceive speech cross-modally. Finally, we exposed two groups of 6-month-old infants to artificially created auditory syllables located between two stereotypical English syllables in the formant space. The auditory syllables followed, equally for both groups, a unimodal statistical distribution, suggestive of a single phoneme category. The visual syllables combined with the auditory syllables, however, were different for the two groups, one group receiving visual stimuli suggestive of two separate phoneme categories, the other receiving visual stimuli suggestive of only one phoneme category. After a short exposure, we observed different learning outcomes for the two groups of infants. The results thus show that visual speech can influence learning of phoneme categories. Altogether, the results demonstrate that complex language learning skills exist from birth. They also suggest a role for the visual component of speech in the learning of phoneme categories.