D O N O T C O Aaron Boes, MD, PhD P Sidney R. Baer Clinical - - PowerPoint PPT Presentation

Network Imaging and Brain Stimulation Aaron Boes, MD, PhD

Sidney R. Baer Clinical Neuroscience Fellow, Berenson Allen Center for Noninvasive Brain Stimulation Pediatric Neurologist, MGH

D O N O T C O P Y

Topics

1) A network model for brain function and dysfunction 1) What can network imaging contribute to brain stimulation? 2) How does brain stimulation modify networks?

D O N O T C O P Y

Disclosure: Off-label uses of TMS will be discussed

D O N O T C O P Y

Mike Fox contributed several slides

D O N O T C O P Y

Please interrupt, ask questions.

D O N O T C O P Y

Topics

1) A network model for brain function and dysfunction 1) What can network imaging contribute to brain stimulation? 2) What can brain stimulation contribute to network imaging?

D O N O T C O P Y

Historical Background: Localization of function

Paul Broca, 1861

D O N O T C O P Y

Localization of functional centers >> Network localization of functions Complex functions arise through interaction among a set of region, each with specialized processing

D O N O T C O P Y

Normal functions result from network-level interactions Dysfunction also due to abnormal network activity

D O N O T C O P Y

Brain disorders will be treated more effectively by identifying and targeting therapy toward dysfunctional networks.

D O N O T C O P Y

Topics

1) A network model for brain function and dysfunction 1) What can network imaging contribute to brain stimulation? 2) What can brain stimulation contribute to network imaging?

D O N O T C O P Y

Major types of network imaging

1) Task-based MRI > identify co-active sites 1) Structural connectivity

• Diffusion tensor imaging

2) Correlated functional measures

• BOLD in fMRI, EEG, MEG

D O N O T C O P Y

• 1
• 0.5

0.5 1 1.5 2 2.5 3 50 100 150 200 250

Open Open Open Open Closed Closed Closed Closed

Open – Closed =

Classical Neuroimaging

% BOLD Change Time (s)

Fox and Raichle (2007) Nat. Rev. Neuro.

Classical functional imaging design

Slide courtesy of Mike Fox

D O N O T C O P Y

Open – Closed =

BOLD Data Is Very “Noisy”

% BOLD Change

Open Open Open Open Closed Closed Closed Closed

• 1
• 0.5

0.5 1 1.5 2 2.5 3 50 100 150 200 250

Time (s)

Fox and Raichle (2007) Nat. Rev. Neuro.

In task-based fMRI bold signal has “noise”

Slide courtesy of Mike Fox

D O N O T C O P Y

BOLD activity oscillates at rest

D O N O T C O P Y

BOLD Data Is Very “Noisy”

% BOLD Change Time (s)

Task-based activations = 1% of brain energy. Majority of energy used to generate spontaneous activity, or “fMRI noise”

Raichle, 2006

D O N O T C O P Y

Spontaneous Fluctuations (“Noise”) in the BOLD Signal

Using BOLD ‘noise’ as signal

D O N O T C O P Y

• 1.5
• 1
• 0.5

0.5 1 1.5 2 50 100 150 200 250 300 Time (sec) % BOLD Change

Spontaneous Fluctuations (“Noise”) in the BOLD Signal

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97 101 105 109

Left Motor Cortex Right Motor Cortex

Using BOLD ‘noise’ as signal

Bharat Biswal et al, 1997

D O N O T C O P Y

1 5 9 13 17 21 25 29 33 37 41 45 49 53 57 61 65 69 73 77 81 85 89 93 97 101 105 109

Left Motor Cortex Right Motor Cortex

• 1.5
• 1
• 0.5

0.5 1 1.5 2 50 100 150 200 250 300 Time (sec) % BOLD Change

Spontaneous Fluctuations are Specifically Correlated

After Bharat Biswal and colleagues (1995) Magnetic Resonance in Medicine

D O N O T C O P Y

Z score, fixed effects, N = 10 Fox and Raichle (2007) Nat. Rev. Neuro.

Generation of Resting State Functional Connectivity Maps

• 1.5
• 1
• 0.5

0.5 1 1.5 2 50 100 150 200 250 300 Time (sec) % BOLD Change

D O N O T C O P Y

• 2.5
• 2
• 1.5
• 1
• 0.5

0.5 1 1.5 2 50 100 150 200 250 300

Time (sec) % BOLD Change

Seed Region in Pcc MPF

Connectivity of site that is active during rest, inactive during tasks

D O N O T C O P Y

• 2.5
• 2
• 1.5
• 1
• 0.5

0.5 1 1.5 2 50 100 150 200 250 300

Time (sec) % BOLD Change

Seed Region in Pcc MPF

D O N O T C O P Y

• 2.5
• 2
• 1.5
• 1
• 0.5

0.5 1 1.5 2 50 100 150 200 250 300

Time (sec) % BOLD Change

Seed Region in Pcc MPF

Default mode network

D O N O T C O P Y

• 2.5
• 2
• 1.5
• 1
• 0.5

0.5 1 1.5 2 50 100 150 200 250 300

Time (sec) % BOLD Change

Seed Region in Pcc MPF IPS Fox et al. (2005) PNAS

Regions activated by tasks Have negative correlation with default mode network

D O N O T C O P Y

Fox et al. (2005) PNAS

Regions activated by tasks Have negative correlation with default mode network

Fox et al. (2005) PNAS

D O N O T C O P Y

DTI Network Rs-fcMRI Network

Honey et al. 2009 PNAS

Resting-state functional connectivity networks correspond to known anatomical paths

D O N O T C O P Y

Exponential Popularity of Rs-fcMRI

Snyder et al. 2012 Neuroimage

Exponential popularity of rs-fcMRI

D O N O T C O P Y

• Understanding disease pathophysiology
• Biomarkers / Diagnosis
• Guiding treatment

Clinical Implications of rs-fcMRI

D O N O T C O P Y

• Understanding disease pathophysiology
• Biomarkers / Diagnosis
• Guiding treatment

Clinical Implications of rs-fcMRI

D O N O T C O P Y

Sample Case – Using rs-fcMRI to gain insight about clinical problem

D O N O T C O P Y

Case

17 yo girl presents with visual hallucinations

D O N O T C O P Y

Case

17 yo girl presents with visual hallucinations

• “Eyes were like a zoom lens, going in and out
• f focus”

D O N O T C O P Y

Case

17 yo girl presents with visual hallucinations

• “Eyes were like a zoom lens, going in and out
• f focus”
• “Scene being drawn in by crayon”

D O N O T C O P Y

Case

17 yo girl presents with visual hallucinations

• “Eyes were like a zoom lens, going in and out
• f focus”
• “Scene being drawn in by crayon”
• Felt as though she was a referee in a soccer

game

D O N O T C O P Y

Case 17 yo girl presents with visual hallucinations

• “Eyes were like a zoom lens, going in and out
• f focus”
• “Scene being drawn in by crayon”
• Felt as though she was a referee in a soccer

game

• Reached for jacket and flowers sprouted from it,

then fell over with straight stems, like popsicle sticks

D O N O T C O P Y

D O N O T C O P Y

Peduncular Hallucinosis (PH)

Predominately visual hallucinations following brainstem

• r thalamus insult

D O N O T C O P Y

Peduncular Hallucinosis (PH)

Where do lesions causing peduncular hallucinosis localize to?

D O N O T C O P Y

Lesion Overlap Analysis Results NN=98

D O N O T C O P Y

Hypothesized mechanism of peduncular hallucinosis

Kazui et al, 2009

Subcortical lesions cause a ‘release’ of visual association cortex (Cogan, 1973; Manford, 1998; Kosslyn, 2001; Kazui, 2009)

D O N O T C O P Y

D O N O T C O P Y

D O N O T C O P Y

Peduncular hallucinosis lesions are anticorrelated to extrastriate visual cortex

NN=98

D O N O T C O P Y

Can this approach be used for other lesion syndromes that have been challenging to localize?

NN=98

D O N O T C O P Y

Boes et al, in prep

D O N O T C O P Y

• Networks associated with focal brain lesions

may serve as targets for repetitive TMS to augment recovery Conclusion: The network effects of focal brain lesions can provide insight about the symptoms

D O N O T C O P Y

Full description of clinical implications of rs-fcMRI beyond scope of talk

D O N O T C O P Y

Disease/Condition References Findings

Alzheimer’s (Allen et al. 2007; Greicius et al. 2004; Li et al. 2002; Supekar et al. 2008; Wang et al. 2006a; Wang et al. 2007; Wang et al. 2006b) Decreased correlations within the default mode network including hippocampi and decreased anticorrelations between the DMN and TPN PIB positive (Hedden et al. 2009; Sheline et al. 2009) Decreased correlations within the default mode network Mild Cognitive Impairment (Li et al. 2002; Sorg et al. 2007) Decreased correlations within the default mode network and decreased anticorrelations between the DMN and TPN Fronto-Temporal Dementia (Seeley et al. 2007a; Seeley et al. 2008) Decreased correlations within the salience network Healthy Aging (Andrews-Hanna et al. 2007; Damoiseaux et al. 2007) Decreased correlations within the default mode network Multiple Sclerosis (De Luca et al. 2005; Lowe et al. 2002) Decreased correlations within the somatomotor network ALS (Mohammadi et al. 2009) Decreased connectivity in DMN and premotor cortex Depression (Anand et al. 2009; Anand et al. 2005a; b; Bluhm et al. 2009a; Greicius et al. 2007) Variable: Decreased connectivity between dACC and limbic regions (amygdala, medial thalamus, pallidostriatum) increased connectivity within the DMN (esp. subgenual prefrontal cortex), decreased connectivity between DMN and caudate Bipolar (Anand et al. 2009) Decreased corticolimbic connectivity PTSD (Bluhm et al. 2009c) Decreased connectivity in the DMN Schizophrenia (Bluhm et al. 2007; Bluhm et al. 2009b; Jafri et al. 2008; Liang et al. 2006; Liu et al. 2006; Liu et al. 2008; Salvador et al. 2007; Whitfield-Gabrieli et al. 2009; Zhou et al. 2007) Variable: Decreased or increased DMN connectivity Schizophrenia 1 relatives (Whitfield-Gabrieli et al. 2009) Increased connectivity in the DMN ADHD (Cao et al. 2006; Castellanos et al. 2008; Tian et al. 2006; Wang et al. 2008; Zang et al. 2007; Zhu et al. 2008; Zhu et al. 2005) Variable: reduced connectivity within the DMN, reduced anticorrelations, increased connectivity in salience Autism (Cherkassky et al. 2006; Kennedy and Courchesne 2008; Monk et al. 2009; Weng et al. 2009) Decreased connectivity within the DMN (although hippocampus is variable and connectivity may be increased in younger patients) Tourette Syndrome (Church et al. 2009) Delayed maturation of task-control and cingulo-opercular networks Epilepsy (Bettus et al. 2009; Lui et al. 2008; Waites et al. 2006; Zhang et

• al. 2009a; Zhang et al. 2009b)

Variable: decreased connectivity in mult. networks including medial temporal lobe, decreased connectivity in DMN with generalized seizure Blindness (Liu et al. 2007; Yu et al. 2008) decreased connectivity within the visual cortices and between visual cortices and somatosensory, frontal motor and temporal multisensory cortices Chronic Pain (Cauda et al. 2009a; Cauda et al. 2009c; Cauda et al. 2009d; Greicius et al. 2008) Variable: Increased/decreased connectivity within the salience network, decreased connectivity in attention networks Neglect (He et al. 2007) Decreased connectivity within the dorsal and ventral attention networks Vegetative State (Boly et al. 2009; Cauda et al. 2009b) Progressively decreased DMN connectivity with progressive states of impaired consciousness

Fox and Greicius (2010) Frontiers Sys Neurosci

D O N O T C O P Y

Topics

1) A network model for brain function and dysfunction 1) What can network imaging contribute to brain stimulation? 2) What can brain stimulation contribute to network imaging?

D O N O T C O P Y

TMS propagates trans-synaptically

D O N O T C O P Y

Nagib et al. Neurosurg clin 2011

D O N O T C O P Y

Cerebral cortex as a portal to the rest of the brain

Cerebral cortex as a portal to connected regions

D O N O T C O P Y

• If NIBS propagates transynaptically, then

connectivity is likely to play a critical role in determining the physiological effect

• Rs-fcMRI is a good technique to assess brain

connectivity

• THUS, rs-fcMRI may be able to predict the

impact and guide NIBS

Train of Logic

D O N O T C O P Y

Rs-fcMRI to target deep epileptic foci

>

• We know that targeting epileptic foci on the

cerebral cortex with 1 Hz rTMS is effective.

D O N O T C O P Y

D O N O T C O P Y

Bernard Chuang and Mo Shafi Annals of Neurology, 2015

D O N O T C O P Y

Can rs-fcMRI be used to improve rTMS therapy in depression?

>

D O N O T C O P Y

FDA approved therapeutic use: High frequency TMS to the DLPFC for medication-refractory depression

D O N O T C O P Y

Herwig et al. 2001 BIOL PSYCHIATRY 50:58–61

Targeting TMS in Depression: The 5 cm method

D O N O T C O P Y

Targeting TMS in Depression: The 5 cm method

Herwig et al. 2001 BIOL PSYCHIATRY 50:58–61 Only hit the “DLPFC” ~40% of the time

D O N O T C O P Y

TMS targets vary in their efficacy

Herbsman et al. 2009

Effective Ineffective

Fitzgerald et al. 2009

42% responders 18% responders

D O N O T C O P Y

Could therapeutic effect of rTMS be related to connectivity to deeper structures?

>

D O N O T C O P Y

Subgenual anterior cingulate cortex: Abnormal in depression

>

Botteron, 2002 Coryell, 2005 Drevets, 1997 Hirayasa, 1999

D O N O T C O P Y

D O N O T C O P Y

Subgenual suppression correlates with antidepressant response

Mayberg 2009 J. Clin. Invest.

SSRI Placebo SNRI ECT DBS

Mayberg 2009 J. Clin. Invest.

D O N O T C O P Y

Fox et al. 2012 Biol Psych.

D O N O T C O P Y

• Avg. 5cm Target

Fitzgerald Target Less Effective 5cm More Effective 5cm

• 0.15
• 0.10
• 0.05

0.00

Subgenual Correlation (r)

P < 0.005

• 0.30
• 0.20
• 0.10

0.00

Subgenual Correlation (r)

P < 5 x 10-8 vs vs

Effective vs. Ineffective TMS Targets

Fox et al. 2012 Biol Psych.

D O N O T C O P Y

Guiding TMS for depression: conclusions

• Effective versus ineffective TMS sites

differ in their functional connectivity, especially with the subgenual cingulate

D O N O T C O P Y

rTMS induces structural change in subgenual anterior cingulate.

Boes et al, abstract submitted.

32 patients. 13 responders (>50% reduction in HamD) + 161.8 mm3, standard error +/- 61.1, t = 2.65, p = 0.02 Non-responders:

• 36.0 mm3, standard error +/-

101.1, t = 0.36, p = 0.73

D O N O T C O P Y

Should target be individualized?

D O N O T C O P Y

Individualized TMS targets for depression

Subgenual Seed Z = 28 X = -4 Z = -10 X = 2 Z = 28 X = -4 R Seed Regions / Seed Maps Group Map Subject 1 Subject 2 15

• 15

0.7

• 0.7

15

• 15

A B C Fox et al. 2012 Neuroimage

D O N O T C O P Y

Day 1 Day 2 Day 1 Subject 1 Subject 2 0.7

• 0.7

A B C Day 2 Subgenual Seed Efficacy- based Seed Map Day 1 Day 2 Day 1 Subject 1 Subject 2 0.7

• 0.7

A B C Day 2

Individual differences in functional connectivity are reproducible across days

Fox et al. 2012 Neuroimage

D O N O T C O P Y

Guiding TMS for depression

• Next step:
• Patients without response to rTMS can

trial individualized targeted approach

D O N O T C O P Y

Is this approach broadly applicable? Can the effects of rTMS be guided by experience in deep brain stimulation, where effective subcortical nodes are known?

D O N O T C O P Y

Disease Invasive (DBS) Noninvasive (TMS, tDCS) Addiction NA DLPFC (laterality unclear) Alzheimer’s Fornix Bilateral DLPFC (+/- parietal, temporal) Anorexia NA, Subgenual L DLPFC Depression Subgenual, VC/VS, NA, MFB, habenula Left DLPFC, R DLPFC Dystonia GPi SMA/ACC, Premotor Epilepsy Thalamus (AN, CM), MTL Active EEG focus Cerebellum Essential Tremor VIM Midline Cerebellum, Lateral Cerebellum, M1 Gait Dysfunction PPN M1 (leg area) Huntington’s GPi SMA Minimally Conscious Thalamus (intralaminar/CL, CM/Pf) R DLPFC, M1 Obsessive Compulsive Disorder VC/VS, NA, ALIC, STN L orbitofrontal, Pre-SMA Pain PAG, Thalamus (VPL/VPM) M1 Parkinson’s STN, GPi M1, SMA Tourette’s Thalamus (CM/Pf), GPi, NA, ALIC SMA

Fox et al. 2014 PNAS

D O N O T C O P Y

Fox et al. 2014 PNAS Invasive and Noninvasive Brain Stimulation Sites are Linked Across 14 Diseases

Addiction Alzheimer’s Anorexia Depression Dystonia Essential Tremor Epilepsy Gait Dysfunction Huntington’s Minimally Conscious OCD Pain Parkinson’s Tourette’s Positive Negative Excitatory Target Inhibitory Target

D O N O T C O P Y

Invasive and Noninvasive Brain Stimulation Sites are Linked Across 14 Diseases

Addiction Alzheimer’s Anorexia Depression Dystonia Essential Tremor Epilepsy Gait Dysfunction Huntington’s Minimally Conscious OCD Pain Parkinson’s Tourette’s Positive Negative Excitatory Target Inhibitory Target

0.02 0.04 0.06 0.08 0.1 DBS Correlation (r) Best Noninvasive Stimulation Site Random Noninvasive Stimulation Sites

P < 0.005 Fox et al. 2014 PNAS

D O N O T C O P Y

Ineffective sites are characterized by an absence of functional connectivity

Parkinson’s Disease Pain Essential Tremor Depression Fox et al. 2014 PNAS

D O N O T C O P Y

The sign of the correlation (positive vs negative) relates to the reported utility of excitatory vs inhibitory stimulation

DBS$ Correla+on$ $ (r)$

DBS$ Correla+on$ $ (r)$

**

• B. All Diseases, Best Sites
• C. All
• A. Parkinson’s Disease

Excitatory Inhibitory M1 Benefit No Benefit SMA No Benefit Benefit

Fox et al. 2014 PNAS

D O N O T C O P Y

The sign of the correlation (positive vs negative) relates to the reported utility of excitatory vs inhibitory stimulation

DBS$ Correla+on$ $ (r)$

DBS$ Correla+on$ $ (r)$

**

• B. All Diseases, Best Sites
• C. All
• A. Parkinson’s Disease

Excitatory Inhibitory M1 Benefit No Benefit SMA No Benefit Benefit

Fox et al. 2014 PNAS

D O N O T C O P Y

Implications…

>

• Network-based rationale for selecting rTMS

targets going forward in clinical trials

D O N O T C O P Y

Topics

1) A network model for brain function and dysfunction 1) What can network imaging contribute to brain stimulation? 2) What can brain stimulation contribute to network imaging?

D O N O T C O P Y

We can target networks using brain stimulation. What happens to the stimulated networks?

D O N O T C O P Y

Modulating the default mode network

D O N O T C O P Y

Eldaief et al. 2012 PNAS

D O N O T C O P Y

D O N O T C O P Y

Changes in connectivity after rTMS to cerebellum

D O N O T C O P Y

So we can alter strength of functional connectivity networks. Does it have functional relevance?

D O N O T C O P Y

Science, 2014

D O N O T C O P Y

D O N O T C O P Y

D O N O T C O P Y

Note the focality of change – 3 mm away = Nonsignificant changes!

D O N O T C O P Y

rTMS-induced functional connectivity correlated with memory performance

D O N O T C O P Y

D O N O T C O P Y

Review

1) A network model for brain function and dysfunction

• Network localization of function and dysfunction

2) What can network imaging contribute to brain stimulation?

• Target deep epileptic foci
• Potentially improve efficacy of depression treatment
• Potentially improve target selection via DBS connectivity

3) How does brain stimulation modify networks?

• Focal alteration in network connectivity with functional

correlates.

D O N O T C O P Y

Review

1) A network model for brain function and dysfunction

• Network localization of function and dysfunction

2) What can network imaging contribute to brain stimulation?

• Target deep epileptic foci
• Potentially improve efficacy of depression treatment
• Potentially improve target selection via DBS connectivity

3) How does brain stimulation modify networks?

• Focal alteration in network connectivity with functional

correlates.

D O N O T C O P Y

Review

1) A network model for brain function and dysfunction

• Network localization of function and dysfunction

2) What can network imaging contribute to brain stimulation?

• Target deep epileptic foci
• Potentially improve efficacy of depression treatment
• Potentially improve target selection via DBS connectivity

3) How does brain stimulation modify networks?

• Focal alteration in network connectivity with functional

correlates.

D O N O T C O P Y

Questions?

D O N O T C O P Y