Brain networks and connectivity in ADHD

• 1. Network connectivity for ADHD
  • 1.1. Frontal-parietal attention network in ADHD
    • 1.1.1. Reduced connectivity within the dorsal frontoparietal executive network
    • 1.1.2. Altered structure of the frontal-parietal attention network
    • 1.1.3. Altered connectivity of the frontal-parietal attention network to other brain networks
    • 1.1.4. Reduced connectivity mPFC-dlPFC predicted later inattention
    • 1.1.5. Reduced connectivity sgACC-dlPFC predicted later anxiety/depression symptoms
  • 1.2. Salience network
  • 1.3. Rich-Club / Hubs Connectivity reduced
  • 1.4. Reduced connectivity in idle mode
  • 1.5. Reduced synchronization of attention networks and sensory cortex areas
  • 1.6. DMN (Default Mode Network)
    • 1.6.1. Reduced deactivation capability of the Default Mode Network (DMN)
    • 1.6.2. Increased connectivity of the Default Mode Network (DMN)
    • 1.6.3. Normalization of the DMN through stimulants
  • 1.7. Thalamic PFC connectivity in ADHD unidirectional
  • 1.8. Increased connectivity between right caudate nucleus and occipital nucleus accumbens
  • 1.9. Impaired signal variability of spontaneous neuronal activity, reduced entropy
  • 1.10. Connectivity and coherence altered in ADHD
  • 1.11. Dopaminergic brain networks and ADHD
    • 1.11.1. Frontal-Striatal Loop
    • 1.11.2. Frontal-cerebellar loop
    • 1.11.3. Frontal-limbic loop
• 2. Brain networks
  • 2.1. Default mode network (DMN)
  • 2.2. Attention networks
    • 2.2.1. Fronto-parietal (attention/executive) network / anterior attention network
    • 2.2.2. Spatial attention network / posterior attention network / dorsal attention network
    • 2.2.3. Ventral attention network / vigilance network
  • 2.3. Salience network
  • 2.4. Visual network
  • 2.5. Auditory network
  • 2.6. Affective network
  • 2.7. Motor networks
  • 2.8. Cerebellar network
  • 2.9. Language network
  • 2.10. Temporal network
  • 2.11. Limbic network
  • 2.12. Cortical network
  • 2.13. Cingulo-Opercular Network
  • 2.14. Memory network
  • 2.15. Subcortical network
• 3. Neurotransmitters and functional connectivity
  • 3.1. Dopamine and connectivity of brain networks
  • 3.2. Noradrenaline and connectivity of brain networks

1. Network connectivity for ADHD¶

A number of studies have analyzed the connectivity of different brain networks in ADHD.
One study compared 4 of these studies and found large inconsistencies. A deviating connectivity between the default mode network and the executive network in ADHD proved to be robust. Furthermore, the consistency of the results was greatest in ADHD-C.¹ It is possible that the various ADHD subtypes exhibit considerable differences in the connectivity of their brain networks.

One study achieved an ADHD diagnostic accuracy of 73 % (specificity 91.6 %, sensitivity 65.5 %) using an analysis of functional connectivity controlled by machine learning.² In practice, these are still not very helpful values for a diagnosis.

1.1. Frontal-parietal attention network in ADHD¶

The frontal-parietal attention network is involved in cognitive processes, especially attention.

1.1.1. Reduced connectivity within the dorsal frontoparietal executive network¶

One study reports reduced connectivity in the dorsal frontoparietal executive network, consisting of

• Right dlPFC
• Posterior parietal cortex

which correlates with the severity of attention problems in ADHD. This correlation was independent of age or gender
Increased connectivity was also associated with increased attention and better accuracy in NoGo tasks.³

One study found no correlates of network connectivity in relation to the symptom of cognitive flexibility that would have revealed differences between autism spectrum disorders, ADHD or non-affected individuals.⁴

1.1.2. Altered structure of the frontal-parietal attention network¶

ADHD-affected children showed in the frontal-parietal attention network compared to non-affected children

• A larger diameter
• A smaller number of leaves,
• A lower tree hierarchy
• A lower kappa

which also correlated with the ADHD symptom score.

• Clustering coefficient and
• Path length

were unchanged.
In children with ADHD, the frontal-parietal attention network may be more decentralized (line-like topology) and less centralized (star-like topology). This appears to result in more random connectivity with impaired global efficiency and network decentralization in ADHD.⁵

1.1.3. Altered connectivity of the frontal-parietal attention network to other brain networks¶

A meta-analysis found that ADHD⁶

• Increased connectivity between
  • Fronto-parietal network and
    • Default Mode Network
    • Affective network
• Reduced connectivity between
  • Fronto-parietal network and
    • Ventral attention network
    • Somatosensory network

One study found reduced structure-function coupling in bundles of connections between hubs (rich clubs?) and peripheral regions in ADHD. This particularly affected connections between the fronto-parietal network and sensory networks. Reduced structure-function coupling correlated with increased ADHD symptoms and increased heterogeneity in neuronal noise variability in brain regions.⁷

1.1.4. Reduced connectivity mPFC-dlPFC predicted later inattention¶

A study of ADHD children with ADHD found that increased connectivity between mPFC and dlPFC at age 7 correlated with decreased attention by age 11.⁸

1.1.5. Reduced connectivity sgACC-dlPFC predicted later anxiety/depression symptoms¶

A study of ADHD children with ADHD found that reduced connectivity between the mood-relevant subgenual anterior cingulate cortex (sgACC) and the dlPFC at age 7 correlated with increased internalizing symptoms (anxiety, depression) up to age 11.⁸

1.2. Salience network¶

One study found deviations in the connectivity of the salience network in ADHD, consisting of³

• Right anterior insula
• Right dorsal anterior cingulate cortex (rdACC)
• Right ventrolateral PFC (rvlPFC).

Another study found increased connectivity within the salience network and decreased connectivity between the DMN and salience network in non-ADHD-affected siblings compared to their ADHD-affected siblings and healthy controls. ADHD sufferers showed increased connectivity between the DMN and task-positive networks compared to their unaffected siblings and healthy controls. ADHD sufferers and their unaffected siblings showed no concordant changes in functional connectivity to healthy control subjects. These results were largely confirmed by complementary pairwise connectomic comparisons. However, the main connectivity differences between the sibling groups could not be replicated in a closely age- and sex-matched subsample (20 subject-sibling pairs and 60 TD).⁹

1.3. Rich-Club / Hubs Connectivity reduced¶

Rich clusters or hubs are clusters of brain regions that are more strongly connected to each other than the average brain region and therefore form the functional backbone of the brain

One study found a reduced density of rich clubs among structural nodes in adults with ADHD, including¹⁰

• Bilateral precuneus
• Insula
• Caudate nucleus
• Left putamen
• Right calcarine sulcus

Furthermore, adults with ADHD were found to have reduced global efficiency compared to non-affected individuals, which may result from a lower density of rich-club connections in ADHD.¹⁰

Another study found deviating asymmetry patterns in ADHD sufferers in connectivity measurements of rich-club connectivity, e.g. more feeder connections. Furthermore, a reduced right asymmetry was found in connectivity measurements of local connections to several peripheral brain regions. Furthermore, abnormal regional asymmetry scores correlated with ADHD symptoms.¹¹

Another study observed that an organization of brain networks around hubs was associated with reduced cognitive impairment under stress. It also showed that the same neurophysiological characteristics of brain connectivity under stress could be associated with quite different problems. There was no 1:1 correlation between brain network patterns and certain symptoms. Children whose brain networks were not organized around hubs showed a wide variance of symptoms in response to stress.¹²

One study described changes in node efficiency in children with ADHD:¹³

• Increase in the visual attention network
• Increase in the dorsal attention network
• Decrease in the somatomotor network
• Acceptance in the default mode network

Impaired functional connectivity and node efficiency correlated significantly with the severity of ADHD symptoms.

1.4. Reduced connectivity in idle mode¶

In ADHD, reduced global connectivity between different brain networks was found in the resting state. This was interpreted as an indication of ADHD as a developmental delay.¹⁴

1.5. Reduced synchronization of attention networks and sensory cortex areas¶

A desynchronization of attention networks and sensory cortex areas was observed in ADHD sufferers who followed a conversation between several people.¹⁵

1.6. DMN (Default Mode Network)¶

As the brain is active even when it is not performing any tasks, a separate network is responsible for managing rest: the Default Mode Network (DMN). It is active when the mind is not involved in another task or is “resting”.¹⁶ The DMN controls things like mind wandering, daydreaming, relaxing and thinking about yourself and your life. Short-term attention deficits are also said to result from the DMN.¹⁷ It follows that the DMN is deactivated when the brain is performing a task.¹⁸ The DMN is not a single anatomical structure, but involves connections between different areas, including

• Anterior mPFC¹⁸¹⁹
• Posterior cingulate cortex¹⁸¹⁹
• Dorsomedial PFC subsystem¹⁸
• Medial temporal lobe subsystem¹⁸ (Precuneus)¹⁹

1.6.1. Reduced deactivation capability of the Default Mode Network (DMN)¶

FMRI studies of the DMN show dysfunction of the DMN in both children and adults with ADHD. ADHD impairs the ability to quickly switch off the DMN and “switch on” the task-positive networks. As a result, ADHD sufferers remain in a certain mental state for longer. To end the persistence, a stronger sensory stimulus is required.²⁰²¹ This inability to deactivate or switch off the DMN is exacerbated by more cognitively demanding tasks.²²

The deactivability of the DMN in ADHD could be increased by high rewards (which, according to our understanding, cause intrinsic interest) in such a way that it corresponded to the deactivability in non-affected individuals. MPH showed the same effect.²³

1.6.2. Increased connectivity of the Default Mode Network (DMN)¶

One study found higher SNAP-IV and SRS total and subscale scores in adolescents with ADHD. Higher SNAP-IV and SRS scores correlated with

• Higher functional connectivity between
  • DMN (ventromedial PFC) and cinguloopercular network (anterior insula)
  • FPN (dorsolateral and prefrontal cortex) and cinguloopercular network
• Less functional connectivity between
  • DMN (posterior cingulate cortex) and frontoparietal network (inferior parietal lobes)
  • DMN (precuneus) and cinguloopercular network (temporoparietal connection).

The authors conclude that social cognition and communication impairments and ADHD may share abnormal functional connectivity in the DMN, frontoparietal network and cinguloopercular network.²⁴

Other studies have also found altered connectivity of the DMN in ADHD.²⁵ Increased correlations of the DMN in the resting state between the posterior cingulate cortex / precuneus and the right middle frontal gyrus were found. These existed equally in adult ADHD sufferers and their first-degree relatives, but not in unaffected controls. The observed connectivity changes correlated with higher sustained attention problems. Furthermore, this brain-based neurocognitive trait dimensionally explained ADHD symptom variability.²⁶

Another study found increased connectivity between the DMN and task-positive networks in ADHD sufferers compared to their unaffected siblings and healthy controls. Non-ADHD-affected siblings showed increased connectivity within the salience network and decreased connectivity between the DMN and salience network, in contrast to their ADHD-affected siblings and healthy controls. ADHD-affected siblings and their unaffected siblings showed no consistent changes in functional connectivity to healthy control subjects. These results were largely confirmed by complementary pairwise connectomic comparisons, but the main connectivity differences between the sibling groups could not be replicated in a closely age- and sex-matched subsample (20 subject-sibling pairs and 60 TD).⁹

One study found strong coherence of the left dorsal anterior cingulate cortex (dACC) with DMN components in children with ADHD. Furthermore, a seed-to-voxel connectivity analysis using the left dorsal anterior cingulate as a seed region found evidence of higher temporal coherence with other neural networks compared to unaffected children. Children with ADHD appear to have a more distributed resting-state connectivity pattern in the DMN and other networks.²⁷

1.6.3. Normalization of the DMN through stimulants¶

In addition, functional imaging studies in ADHD sufferers not only show differences in DMN regions compared to healthy controls, but also a return to normal functions during treatment with stimulants (here: methylphenidate).^28293031 The more intrinsically motivated attentional control in ADHD means that attention and its controllability is just as high as in non-affected individuals when there is a correspondingly high (intrinsic) interest and only deviates from the attention of non-affected individuals when there is a lower (intrinsic) interest. This is controlled by the DMN. Stimulants are able to align the attention control of ADHD sufferers with that of non-affected persons in the absence of (intrinsic) interest.³²

1.7. Thalamic PFC connectivity in ADHD unidirectional¶

One study found a unidirectional connection from the cortex to the thalamus in the alpha, beta and gamma bands in ADHD sufferers, while non-affected people showed bidirectional connectivity between the cortex and thalamus.³³

1.8. Increased connectivity between right caudate nucleus and occipital nucleus accumbens¶

Functional connectivity between the right parietal caudate and the nucleus accumbens correlates with ADHD.³⁴

1.9. Impaired signal variability of spontaneous neuronal activity, reduced entropy¶

One study reported reduced brain signal variability and reduced multiscale entropy in children with ADHD in the higher-order functional networks as well as in the primary brain networks, e.g. in the default mode, frontoparietal network, attention network and visual network. The deviations in reduced brain signal variability correlated negatively with ADHD symptoms in the frontoparietal network and negatively with reaction time variability in the frontoparietal, default mode, somatomotor and attention networks.³⁵

Another study found a reduced amplitude of multiscale entropy in ADHD. The absolute spectral power showed higher variability values in ADHD. Multiscale entropy changes with age and increases with increasing number of scales. ADHD appears to have increased EEG variability and lower EEG complexity.³⁶

A reversal of entropy in various brain networks has been described from high to low stress.³⁷

1.10. Connectivity and coherence altered in ADHD¶

Several studies report interhemispheric connectivity changes in ADHD:³⁸

• Reduced interhemispheric coherence in the delta band in frontal brain regions³⁹
• Increased coherence in the theta band in posterior regions (only with eyes open)³⁹
• Increased coherence in the theta band in central areas³⁹

In one study, methylphenidate normalized reduced global connectivity in ADHD 400-700 ms after a stimulus and reduced an increase in network disconnection 100-400 ms after the stimulus. These global changes caused by methylphenidate occurred mainly in the task-relevant frontal and parietal regions and were more significant and lasting than in the non-treated comparison subjects. The results of the study indicate that methylphenidate corrects impaired network flexibility in ADHD.⁴⁰

Several studies have analyzed the peculiarities of communication and connection within the brain in ADHD by measuring EEG and QEEG values.

Measuring the synchronization or coherence of the brain’s EEG in the resting state is a common method for measuring cortical functional connectivity.⁴¹

In ADHD, there is a clear pattern of reduced connectivity between anterior (front) and posterior (back) brain regions.⁴²

There is also reduced connectivity in the frontostriatal connections in ADHD.⁴³⁴⁴

Adults with ADHD showed reduced activation of the bilateral inferior PFC, caudate and thalamus during stop tasks and cognitive switch tasks, as well as in the left parietal lobe (only) during the switch task. The more severe the ADHD behavioral symptoms were, the lower the increase in activation of similar regions in fronto-striatal, parietal and cerebellar brain areas. During stop tasks, ADHD sufferers showed reduced inter-regional functional connectivity between right anterior frontal, fronto-striatal and fronto-parietal neural networks.⁴⁴

Another study found a lack of connectivity between the brain hemispheres in ADHD at rest and a stimulus-induced state of hyperconnectivity within and between frontal hemispheres. According to the study, ADHD is characterized by altered functional connectivity, particularly between frontal regions.⁴⁵

At shorter distances between the measuring electrodes, ADHD children showed intrahemispheric coherence in the theta band and reduced lateral differences in the theta and alpha bands. At larger electrode distances, ADHD children showed lower intrahemispheric coherence in the alpha band than non-affected children. In the frontal brain area, ADHD-affected children had increased interhemispheric coherence in the delta and theta bands and decreased interhemispheric coherence in the alpha band. Coherence in the alpha band was reduced in the temporal brain area. Coherence in the theta band was increased in central / parietal / occipital brain regions.
ADHD sufferers of the mixed type had greater intrahemispheric theta and beta coherence than ADHD sufferers of the purely inattentive subtype. Frontally, ADHD sufferers of the mixed type have increased interhemispheric coherence in the delta and theta bands compared to ADHD sufferers of the purely inattentive subtype (ADHD-I). In ADHD-C sufferers, coherence in the beta band is increased in central / parietal / occipital brain regions.⁴⁶
Another study used EEG signals, particularly in the beta band, to find evidence of altered effective connectivity in brain networks in children with ADHD.⁴⁷

ADHD sufferers showed increased coherence in the lower alpha band (8 Hz) and decreased coherence in the upper alpha band (10-11 Hz). The increase in coherence at 8 Hz in ADHD sufferers and at 2 to 6 Hz in the control group was independent of stimulus presentation. In response to visual stimulation, ADHD sufferers showed reduced evoked power and increased frontal coherence. Non-affected persons only differed from ADHD patients on medication in the coherence level at 8 Hz. The results indicate a persistent lack of connectivity and stimulus-induced overconnectivity within and between frontal hemispheres in ADHD.⁴⁵

One study identified four corresponding patterns of dynamic functional brain connectivity (dFC) for different dimensions of behavioral or cognitive functions:⁴⁸

• Dimension inattention/hyperactivity
  • positively associated with dFC within Default Mode Network (DMN)
  • negatively associated with dFC between DMN and sensorimotor network (SMN)
• Dimension somatization
  • positively associated with dFC within the DMN and the SMN
• Inhibition and flexibility dimension
  • positively associated with dFC within DMN
  • positively associated with dFC between DMN and SMN
  • negatively associated with dFC between DMN and fronto-parietal network
• Fluency and memory dimension
  • positively associated with dFC within the DMN
  • positively associated with dFC between DMN and SMN
  • negatively associated with dFC between DMN and fronto-parietal network
    The cognitive functions of inhibition and flexibility mediated the relationship between brain dynamics and the behavioral manifestations of inattention and hyperactivity.

A study of adults with ADHD found widespread intrahemispheric hypoconnectivity in networks that included:⁴⁹

• subcortical structures
• cortical structures
• dorsal attention network
• ventral attention networks
• visual systems
• somatomotor systems

In addition, hypo-connectivity in a predominantly left-frontal network correlated with increased commission errors in CPT-2.⁴⁹

In addition to reduced functional connectivity (especially in visual and sensorimotor brain networks), increased connectivity of peripheral networks was found.⁵⁰

1.11. Dopaminergic brain networks and ADHD¶

Three brain networks and their influence on ADHD are described.⁵¹⁵²⁵³

Altered dopaminergic function cannot adequately modulate non-dopaminergic signaling (mainly glutamate and GABA).⁵⁴

1.11.1. Frontal-Striatal Loop¶

“Cold” executive functions that concern the “what”.⁵⁵

Correlates with problems with

• Response inhibition
• Distractibility
• Working memory
  • Organization
  • Planning

Mesocortical dopamine system:⁵⁴

• Attention deficits
  • Inadequate orientation reactions
  • Impaired saccadic eye movements
  • Poorer attention reactions to a target
• Poor behavioral planning (poor executive functions)

1.11.2. Frontal-cerebellar loop¶

“Hot” executive functions that concern the “when”.⁵⁵

Correlates with problems with

• Motor coordination
• Timing and timeliness of behavior

Nigrostriatal dopamine system:⁵⁴

• Disturbed modulation of motor functions
• Impairment of non-declarative habituation learning and memory.
• These lead to obvious developmental delays

1.11.3. Frontal-limbic loop¶

Executive functions that concern the “why”

Correlates with problems with

• Emotional dysregulation
• Motivation problems
• Hyperactivity/impulsivity
  • Hyperactivity, especially in new situations
• Susceptibility to aggression

Mesolimbic loop:⁵⁴

• Modified reinforcement of the behavioral reinforcement
• Inadequate extinction of a previously reinforced behavior
• Problems with sustained attention
• Increased behavioral variability
• Inhibition problems with the “inhibition” of reactions (“disinhibition”)

2. Brain networks¶

The specialist literature does not use standardized terms for the various brain networks
The following diagram attempts to illustrate the brain networks.

2.1. Default mode network (DMN)¶

• Active when the individual is awake and at rest
• Preferably active when focusing on internal-oriented tasks such as daydreaming, imagining the future, recapitulating memories, Theory of Mind (ToM).
• Not active when other brain networks focus on external (visual) stimuli
  • DMN can be deactivated tonically⁵⁶ and phasically⁵⁷.²³
• The DMN is not a single anatomical structure, but comprises connections between different areas, including¹⁸
  • Anterior medial PFC
  • Posterior cingulate cortex
  • Dorsomedial PFC subsystem
  • Medial temporal lobe subsystem.
• Connected to⁵⁸
  • Ventromedial PFC
  • Dorsomedial PFC
  • Posterior cingulum
  • Retrosplenial cortex
  • Inferior parietal lobe
  • Lateral temporal cortex
  • Hippocampus

A posterior default mode network was also named.

2.2. Attention networks¶

2.2.1. Fronto-parietal (attention/executive) network / anterior attention network¶

• Probable synonyms:

  • Executive network
    The executive network is said to be primarily associated with impulsivity⁵⁹

    • Left
      • Left-wing executive network
      • Left fronto-parietal network
    • Right
      • Right-wing executive network

  • Task control network

• Initiates and modulates cognitive control⁶⁰
  • Inhibition
  • Task change
  • Conflict resolution
  • Allocation of resources
  • Planning
  • Selective discovery of sensory and semantic events.⁶¹
• Components³
  • Right dlPFC
  • Posterior parietal cortex

2.2.2. Spatial attention network / posterior attention network / dorsal attention network¶

Control of the spatial orientation of attention to sensory stimuli.

• Voluntary use of attention and reorientation to unexpected events
• Dopaminergic

The attention shift consists of three sub-processes:⁶²

• Detachment of the focus of attention (“disengagement”)
• Shifting the focus of attention (“shifting”)
• Focusing on the new object of attention (“engagement”).

Three areas of the brain are apparently involved in the spatial attention network:⁶³

• Posterior parietal cortex
  • Detachment of the focus of attention (“disengagement”)
• Superior colliculus
  • Shifting the focus of attention (“shifting”)
• Lateral pulvinar of the thalamus
  • Focusing on the new object of attention (“engagement”)

2.2.3. Ventral attention network / vigilance network¶

Generation and maintenance of a basic activation level or a basic readiness to react.⁶⁴

• Reacts when behaviorally relevant stimuli occur unexpectedly
• Predominantly subcortical⁶²
  • Mesencephalic structures
  • Right hemispheric cortical areas
    • Right PFC
• Noradrenergic⁶⁰

2.3. Salience network¶

• Monitors the meaning of external inputs and internal brain events
• Components³
  • Right anterior insula
  • Right dorsal anterior cingulate cortex (rdACC)
  • Right ventrolateral PFC (rvlPFC)

2.4. Visual network¶

• Lateral visual network
  • Involved in complex emotional stimuli

2.5. Auditory network¶

2.6. Affective network¶

2.7. Motor networks¶

• Sensorimotor network
  • This is primarily associated with inattention⁵⁹
• Somatomotor network
• SM mouth network
• SM manual network

2.8. Cerebellar network¶

2.9. Language network¶

2.10. Temporal network¶

2.11. Limbic network¶

2.12. Cortical network¶

The cortical network is described as a network of⁶⁵

• DlPFC
• Dorsomedial PFC
• Anterior cingulate cortex (ACC)
• Inferior frontal gyrus
• Insula
• Parietal regions.

2.13. Cingulo-Opercular Network¶

2.14. Memory network¶

2.15. Subcortical network¶

3. Neurotransmitters and functional connectivity¶

3.1. Dopamine and connectivity of brain networks¶

Dopamine appears to specifically influence the functional connectivity of certain brain regions. According to one study, amisulpride (a D2/D3 antagonist) increased functional connectivity from the putamen to the precuneus and from the ventral striatum to the precentral gyrus. L-DOPA (a dopamine precursor) increased functional connectivity from the VTA to the insula/operculum and between the ventral striatum and vlPFC and decreased functional connectivity between the ventral striatum and dorsal caudate with the medial PFC.⁶⁶

3.2. Noradrenaline and connectivity of brain networks¶

The pupil diameter at rest, which correlates with tonic noradrenal firing, also reflects the connectivity between frontoparietal, striatal and thalamic brain regions.⁶⁷