Yun Hseng Liao v. Maurice Junious

FOR PUBLICATION
UNITED STATES COURT OF APPEALS
FOR THE NINTH CIRCUIT
YUN HSENG LIAO,
Petitioner-Appellant,
v.
MAURICE JUNIOUS,
Respondent-Appellee.
No. 14-55897
D.C. No.
2:10-cv-05691-JGB-JCG
OPINION
Appeal from the United States District Court
for the Central District of California
Jesus G. Bernal, District Judge, Presiding
Argued and Submitted
October 23, 2015—Pasadena, California
Filed January 29, 2016
Before: Harry Pregerson and Stephen S. Trott, Circuit
Judges and William H. Stafford,* Senior District Judge.
Opinion by Judge Trott
* The Honorable William H. Stafford, Jr., Senior District Judge for the
U.S. District Court for the Northern District of Florida, sitting by
designation.

-- 1 of 35 --

LIAO V. JUNIOUS 2
SUMMARY**
Habeas Corpus
The panel reversed the district court’s denial of California
state prisoner Yun Hseng Liao’s habeas corpus petition
challenging his conviction for assaulting and attempting with
premeditation to kill his ex-girlfriend’s teenage son, and
remanded.
Liao’s unsuccessful defense was that the incident
happened while he was in a state of unconsciousness during
an episode of sleepwalking, and thus, that he lacked the intent
required for the crimes with which he was charged.
During further proceedings on Liao’s ineffective
assistance claim after newly discovered evidence revealed a
significant lapse on trial counsel’s part, the Superior Court
concluded that trial counsel’s performance had been
constitutionally defective by failing to secure medical
evidence to support Liao’s primary expert’s sleepwalking
opinion, but that counsel’s failure had not been prejudicial.
The panel concluded that the Superior Court’s decision
that Liao suffered no prejudice was based on an unreasonable
determination of the facts and was objectively unreasonable
in its application of clearly established Federal constitutional
law.
** This summary constitutes no part of the opinion of the court. It has
been prepared by court staff for the convenience of the reader.

-- 2 of 35 --

LIAO V. JUNIOUS 3
COUNSEL
Dennis A. Fischer (argued) and John M. Bishop, Law office
of Dennis A. Fischer, Santa Monica, California, for
Petitioner-Appellant.
Ryan M. Smith (argued), Deputy Attorney General, Kamala
D. Harris, Attorney General of California, Gerald A. Engler,
Chief Assistant Attorney General, Lance E. Winters, Senior
Assistant Attorney General, and Kenneth C. Bryne,
Supervising Deputy Attorney General, Los Angeles,
California, for Respondent-Appellee.
OPINION
TROTT, Senior Circuit Judge:
On June 16, 2003, a jury in the Superior Court of Los
Angeles County, California convicted appellant Yun Liao of
assaulting and attempting with premeditation to kill Henry
Chen, his ex-girlfriend Li’s teenage son. Liao admittedly hit
Chen three times on the head with a hammer at 4:00 a.m
while Chen was asleep. Liao’s unsuccessful defense was that
the incident happened while he was in a state of
unconsciousness during an episode of sleepwalking, and thus,
that he lacked the intent required for the crimes with which he
was charged. The court sentenced him to life in prison with
the possibility of parole, plus four years. Twelve years later,
he has served his time and is out of prison on parole.
Liao’s appeal as well as his pursuit of state habeas corpus
relief failed, but because of newly discovered evidence
revealing a significant lapse on his trial counsel’s part, the

-- 3 of 35 --

LIAO V. JUNIOUS 4
California Court of Appeal returned his case to the Superior
Court for further proceedings on his claim of ineffectiveness
of counsel. The Court of Appeal ordered the California
Department of Corrections and Rehabilitation to show cause
why Liao’s conviction should not be set aside. After a
hearing, the Superior Court concluded in a decision spoken
from the bench that trial counsel’s performance had indeed
been constitutionally defective by failing to secure medical
evidence to support his primary expert’s sleepwalking
opinion – a conclusion with which the prosecution agreed.
The Superior Court found, however, that counsel’s failure had
not been prejudicial.
After unsuccessful attempts in state court to overturn the
Superior Court’s decision, Liao filed a petition for a writ of
habeas corpus in the Central District of California, alleging a
violation of his Sixth Amendment right to effective assistance
of counsel. In an order accepting the report and
recommendation of a magistrate judge agreeing with the
Superior Court, the district court denied Liao’s petition. He
appeals.
We have jurisdiction over this timely appeal pursuant to
28 U.S.C. §§ 1291 and 2253. Because we conclude that the
Superior Court’s decision was (1) based on an unreasonable
determination of the facts, and (2) objectively unreasonable
in its application of clearly established Federal constitutional
law, we reverse.

-- 4 of 35 --

LIAO V. JUNIOUS 5
I
Facts1
At about 4:00 a.m. on August 4, 2002, Henry Chen was
awakened by the sensation of three blows on his head. They
were inflicted by a household hammer, which Chen had left
on the floor of his room. He covered his head with his hands,
and in the dark sought his assailant, whom he pushed to the
ground. Chen then saw that it was Liao, and he asked Liao
what he was doing. Liao did not reply. Chen ran into his
mother Li’s room, passing his younger brother Danny, and
telling her Liao had hit him. She covered his head, and asked
Liao to call the police. He stated he would go to jail, but after
several requests, he made the call. Li asked Liao why he hit
Chen. After repeating, “Why did I do it?” Liao replied that
he had been dreaming someone was hitting him, and he had
fought back.
As Chen walked outside to meet arriving paramedics,
Liao told him to say he had fallen down the stairs, because
otherwise he would go to jail.
Chen suffered three scalp lacerations, each over an inch
long, which were closed by staples, as well as a cut to his ear.
He also suffered a concussion, and both of his hands were
fractured. He remained at the hospital for about six hours.
The attending physician opined that Chen had received a
series of glancing but direct blows to the head. There was no
skull or brain damage. At the time of trial, Chen still
1 We borrow these facts without attribution from the California Court of
Appeal’s unpublished decision. People v. Yun Hseng Liao, Second
Appellate District, No. B170596.

-- 5 of 35 --

LIAO V. JUNIOUS 6
experienced headaches and dizziness, as well as some pain in
his hands.
Chen had known Liao for five years, during which time
Liao had been Li’s boyfriend, and had lived with her and her
sons for four years. In that time, Liao had never previously
struck Chen, although, according [to] Chen, he had hit Danny
once, and Li twice (out of Chen’s presence). Liao had shown
no animosity toward Chen, who believed Liao had attacked
him out of anger at Li.
On the night of the incident, Chen testified that Liao and
Li had had an argument in her room. Li then stopped a fight
between the brothers, and Liao told Chen not to make Li
angry. Before Chen went to bed around midnight, he saw
Liao, smoking a cigarette and staring out the sliding window
by the balcony. Chen testified that Liao smoked when he was
“stressed.” Danny went to sleep on the living room couch.
When awakened by the blows, Chen, who had been sleeping
face down, threw Liao off, and then saw him. Liao seemed
shocked, and stared at Chen, mouth open. Liao was holding
the hammer, raised, while leaning against a wall about seven
feet from Chen. Chen then asked Liao what he was doing,
and received no reply.
Danny had gone to sleep about 9:00 p.m. He was
awakened by the sound of three blows, like a hammer striking
a nail. He saw his brother bleeding profusely. Li asked Liao
to call the police. The first time he refused, saying that if he
went to jail this time it would cost him a lot of money to get
out. After a second request he did call, and he accompanied
Li to the hospital.

-- 6 of 35 --

LIAO V. JUNIOUS 7
II
A.
Counsel’s Error
In preparation for Liao’s trial, his counsel hired Dr. Clete
Kushida, a Director of the Stanford University Center for
Human Research and a board certified physician at Stanford’s
Sleep Disorders Clinic. Dr. Kushida is also on the academic
faculty of the Stanford School of Medicine. After reviewing
the facts and circumstances of Liao’s unusual pre-dawn
behavior, Dr. Kushida recommended that Liao undergo a
medical examination and a “sleep study,” formally known as
polysomnogram.
Dr. Kushida put his recommendations in a letter dated
April 8, 2003. In the letter, Dr. Kushida said that “further
evaluation of Mr. Liao is warranted, with a consideration of
a sleepwalking diagnosis.” The inquiry would include “[a]n
evaluation conducted by a sleep specialist. This would entail
a review of Mr. Liao’s medical history and a physical
evaluation.” Dr. Kushida also recommended “[a]n overnight
polysomnogram (sleep study) conducted by an experienced
polysomnographic technologist, and reviewed by a sleep
specialist . . . an important component of the evaluation of an
individual with a diagnosis of sleepwalking.”
Because Liao was in custody, counsel filed a request with
the Superior Court for authorization to conduct the
procedures recommended by Dr. Kushida. On April 10,
2003, a court commissioner – not the trial judge – denied the
motion without prejudice. On April 25, 2003 counsel
promptly renewed his motion, supplementing it with a second

-- 7 of 35 --

LIAO V. JUNIOUS 8
letter from Dr. Kushida dated April 22, 2003, reiterating the
necessity of a sleep study. We will discuss the content of this
second letter in more detail in part B. of this opinion. The
commissioner took the matter under submission. When
counsel’s associate later called the court to inquire about the
status of the request, a court clerk erroneously told him that
the motion had been denied when in fact it had been granted
on May 1, 2003. On the commissioner’s Order, he wrote, “1.
To be completed prior to trial date - 2. Not to exceed $2500,”
followed by his signature. The approving Order, prepared by
counsel and signed by the court commissioner, lay
undiscovered in the court’s file until Liao’s conviction was on
appeal.
Instead of conducting any further inquiry into the status
of his motion, counsel proceeded to trial without the benefit
of the medical examination and study for which Dr. Kushida
had asked. During the trial, the absence of a sleep study
turned out to be the Achilles heel of Liao’s defense.
At all stages of the proceedings, California has conceded
that counsel’s failure to verify what the court clerk told his
associate over the phone amounted to constitutionally
ineffective assistance under Strickland v. Washington,
466 U.S. 668 (1984).
B.
The Effect at Liao’s Trial
of the Absence of a Sleep Study
To establish Liao’s sleepwalking / lack of criminal intent
defense, counsel called Dr. Kushida as an expert witness.
When Dr. Kushida took the stand however, he had none of

-- 8 of 35 --

LIAO V. JUNIOUS 9
the material referenced in his letter to rely on or to support his
opinion. He testified that because he did not have what he
had asked for, a sleep study and physical examination, he
could not diagnose Liao as a sleepwalker, only render an
opinion that he suffered from that condition.
The absence of this information enabled the prosecutor to
discredit Dr. Kushida’s testimony on cross-examination and
to render his opinion suspect. The prosecutor’s first question
was whether Dr. Kushida had interviewed the defendant. His
answer was no. The second question was whether he had
“conducted any sleep studies of the defendant,” to which Dr.
Kushida gave the same negative answer. The third question
was a statement: “So your opinion is based on – well,
obviously your opinion is not based on anything that the
defendant has told you or anything that you have observed in
studying his sleep patterns, correct?” Answer, “That is
correct.”
Over and over the prosecutor returned to the absence of
a sleep study and physical examination to support Dr.
Kushida’s testimony—an absence caused by counsel’s error.
On re-cross, she effectively hammered Dr. Kushida with his
second pre-trial letter submitted to the court commissioner in
support of counsel’s motion, using his own words
recommending an examination and a sleep study.
Q: The Prosecutor A: Dr. Kushida
Q Doctor, directing your attention to your
letter to Mr. Donoghue, or his associate, dated
April the 22nd. I think you have a copy in
your hand?

-- 9 of 35 --

LIAO V. JUNIOUS 10
A Yes, I do.
Q Do you state in that letter that
sleepwalking is suspected in Mr. Liao’s case
because of the following?
A Yes, I do.
Q Okay. Do you also state in that letter that
the above points raise clinical suspicion that
Mr. Liao’s behavior during the episode in
question is compatible with sleepwalking?
A Yes.
Q Okay. Do you also say in that letter that
further evaluation is warranted?
A Yes.
Q Do you continue to explain, specifically
on page 2 of your letter that an overnight
polysonogram [sic], a sleep study, is
warranted to further evaluate Mr. Liao’s
preliminary sleepwalking diagnosis?
A Yes.
Q And then do you give the rationale for a
sleep study in Mr. Liao’s case?
A Yes, I do.

-- 10 of 35 --

LIAO V. JUNIOUS 11
Q And there are three different rationales,
correct?
A Yes.
Q You also state that it’s – a sleep study is
an objective test that is used to assess patients
with sleep disorders, correct?
A Yes, I do.
Q And the last sentence you say that in the
case of Mr. Liao the sleep study is an
important component in his evaluation for a
diagnosis of sleepwalking, correct?
A That is correct.
. . . .
Q During my cross-examination did you say
that – did you testify that you could not
conclude that the defendant was a
sleepwalker, just that he may be a
sleepwalker?
A What I said was I could not make an
official diagnosis because official diagnosis
depends on actually clinically examining the
patient.
Q But you said that it was your opinion that
he may be a sleepwalker?

-- 11 of 35 --

LIAO V. JUNIOUS 12
A Yes.
Q Okay. But – and it was also your opinion
that you needed to do a sleep study because it
was important to the diagnosis of him being a
sleepwalker, correct?
A At that time, yes.
. . . .
Q Well, then why did you want to do a sleep
study as you stated in you letter and as I
thought you stated this morning?
A To rule out other potential causes that –
the main reason is that there are things that
can mimic sleepwalking and that is nocturnal
seizures or epilepsy. That’s the first point.
And that can really be confused with
sleepwalking and that’s easily treated by
putting the person on anticonvulsants. That’s
the first point.
The second point is looking at markers on
the sleep study because there are some
elements on the sleep study that you can find
that are non specific markers, you know, that
indicate that the person might have
sleepwalking. The third reason is to rule out
other sleep disorders such as sleep apnea or
periodic limb movements that can fragment
the sleep and trigger off a sleepwalking
episode. So that’s the reason. It’s more to

-- 12 of 35 --

LIAO V. JUNIOUS 13
look at the etiology of the condition, the cause
of the condition.
. . . .
Q And now you are saying that your opinion
is that he is a sleepwalker, correct?
A Correct.
Q That indicates to me that there has been a
change in opinion since you have been on the
witness stand from this morning.
A I would have to go back, look at exactly
what I said. But it’s my opinion that he is a
sleepwalker. At that time maybe perhaps I
was thinking about the actual diagnosis. And
just to reiterate, to actually make a diagnosis,
I would have to actually evaluate the patient.
If you were to ask me what my opinion is, my
opinion is he is a sleepwalker.
. . . .
The Court: Doctor, tell us the difference
between your opinion and a medical
diagnosis.
The Witness: Yes. For medical diagnosis I
would actually have to see the patient, and,
you know, lay hands on the patient, physically
examine the patient. In terms of an opinion,
it’s based on material that I acquired about the

-- 13 of 35 --

LIAO V. JUNIOUS 14
patient to actually make my opinion regarding
diagnosis. But I can’t actually say the word I
diagnose this patient as having a condition
without actually talking with the patient. But
to clarify, you know, based on the material
that I have received, you know, it is my
opinion that he is a sleepwalker. It’s just that
I can’t actually say I diagnosed the patient as
being a sleepwalker because I haven’t actually
physically evaluated the patient.
C.
Rebuttal
In rebuttal to Dr. Kushida’s vulnerable opinion testimony,
the prosecution called Dr. Kaushal Sharma, a physician board
certified in psychiatry. Dr. Sharma’s expertise was in
applying “knowledge and information about a suspected
sleepwalker to the forensic issue.” He admitted he had “not
run any sleep lab, therefore I would not call myself
exclusively an expert in the field of sleepwalking.” The
prosecution used Dr. Sharma to counter Dr. Kushida’s
opinion. He did, using the absence of a sleep study as his
weapon.
Q: Prosecutor A: Dr. Sharma
A I was given the task of looking over the
documents you described and helping,
initially, understanding a little bit more about
the sleep disorders including sleepwalking.
And I was given the task of looking at the
report or letters of two of my professional

-- 14 of 35 --

LIAO V. JUNIOUS 15
colleagues and seeing if medically what they
had described based on the information they
had and whatever else they may have done
was the diagnosis consistent with the
information they had, and to tell you if I
agreed, in what aspect, and if I did not, in
what aspect.
Q And did – in that process have you also
formed your own opinion with regard to the
defendant’s mental status?
A Yes.
Q Let me just directly jump into whether or
not you believe the defendant is a sleepwalker
or has sleep disorders.
A Sleep disorders range from having
difficulty falling asleep. He may have that
problem in the jail. I don’t know for a fact.
But specifically about sleepwalking I believe
the information given to me is insufficient to
prove that he suffers from sleepwalking
disorder.
Q And any information – well, why is that?
A Any confirmed diagnosis of
somnambulism, s-o-m-n-a-b-u-l-i-s-m [sic],
which is just the fancy term for sleepwalking,
requires that the person be subjected to sleep
lab tests where electrodes are placed on the
person’s brain and the brain’s activity, as well

-- 15 of 35 --

LIAO V. JUNIOUS 16
as the activity of the eyes, because they move
at a rapid pace in certain phases of the sleep,
is detected. And then it’s seen if the person is
doing activity which is inconsistent with the
normal average sleep.
To the best of my knowledge the
defendant in this case was not given any such
sleep lab test. So that’s the one problem I
have and therefore I believe that there is
insufficient data.
On cross-examination, Dr. Sharma stuck to his guns.
Q: Defense Attorney A: Dr. Sharma
Q Now your opinion there is insufficient
information to determine a diagnosis whether
or not the defendant suffered from
sleepwalking. Is – what about as to an
opinion, a medical opinion as compared to a
medical diagnosis, is there a difference?
A Well, diagnosis – well, in many ways they
are. In the context, they are the same.
Diagnosis is an opinion.
Q Then are all opinions diagnoses?
A No, they are not.
Q So there is a difference?
A Yes.

-- 16 of 35 --

LIAO V. JUNIOUS 17
Q So it’s possible for a doctor to reach – to
form an opinion without yet having a
diagnosis, is that correct?
A Sure. A doctor may have opinion that
there is no diagnosis.
Q Is it your opinion that a – it’s your
opinion, I believe you testified, that a
diagnosis requires that the person be subjected
to a sleep lab test, correct?
A In this context, yes.
D.
Jury Argument
Having set up her summation with the precision of a
surgeon, the prosecutor belittled and mocked Dr. Kushida’s
testimony in final argument.
(Prosecutor) Dr. Kushida’s opinion, we have
three different things with Dr. Kushida. I
couldn’t believe the way he testified on the
stand, frankly, ladies and gentlemen. He says
that in his letters to the defense attorney
sleepwalking is suspected, you know, et
cetera, et cetera. This raises a clinical
suspicion. But the sleep study is very
important. The sleep study is very important.
Then when he comes to the stand and he
testifies in court he starts out and he says,

-- 17 of 35 --

LIAO V. JUNIOUS 18
yeah, the defendant may be a sleepwalker.
But I mean [sic] need to do the sleep study in
order to diagnose him or in order to say that
he is a sleepwalker. This happened in direct
examination, in my cross-examination. But
when we come to the defendant’s or the
defense attorney’s redirect examination, what
does he say, he changes his testimony, right
here in front of us, in front our eyes. He says
the defendant is a sleepwalker. Oh, yeah,
with medical certainty. That means a doctor’s
guarantee, a stamp of medical approval. How
can you tell me that’s not a diagnosis. He
says with medical certainty. And sleep study
is not required contrary to what he said. Then
– and then let’s look – so, you know, these are
the doctors the defense is putting up and
asking you to be convinced. That’s
unreasonable, ladies and gentlemen. These
doctors have changed their testimony.
With regard to Dr. Sharma’s opinion, Dr.
Sharma is the only consistent doctor. His
duty was to review Dr. Kushida’s’s [sic]
opinions and letters, review Dr. Vicary’s
opinions and letters and his findings in that 8-
page report, review the preliminary hearing
transcript, the police reports, the evidence, so
on so forth, and tell us whether or not he
thought – whether or not he agreed with them.
And that was a major risk on my part because
he could have agreed with him. But you
know what, it defies these two doctors, defies
common sense. And Dr. Sharma told us that.

-- 18 of 35 --

LIAO V. JUNIOUS 19
He said there is insufficient evidence that the
defendant is a sleepwalker. He was very – I
think very professional about that. He could
have said, you know, there is no evidence that
the defendant is a sleepwalker or I don’t think
the defendant is a sleepwalker for X, Y, and Z
reasons. But he tells us there is insufficient
evidence. Why? Because there was no sleep
study that was done, which is important.
III
The Sleep Study
The effect on Liao’s trial of counsel’s failure to secure an
examination and a sleep study was litigated on remand in an
evidentiary hearing in the Superior Court. Before the
hearing, Liao finally received his sleep study, conducted in
two phases by Dr. Milton Erman, a distinguished fellow of
both the American Psychiatric Association and the American
Academy of Sleep Medicine. Dr. Erman completed his
psychiatric residency in Boston at Massachusetts General
Hospital and a four-year fellowship at Harvard Medical
School before going on to practice his specialty. In arriving
at his diagnosis regarding Liao, Dr. Erman consulted an
expert from Stanford, Dr. Guilleminault. Dr. Guilleminault
is a world-renowned specialist in sleep disorders who has
created a proprietary computer scoring technique called
“power spectral analysis” which analyzes objective data
obtained from a patient during sleep studies.
Based on the objective and subjective data collected
during this detailed process, the doctors diagnosed Liao as a
somnambulist, or a sleepwalker. They supported their

-- 19 of 35 --

LIAO V. JUNIOUS 20
collective opinion with their findings of sleep apnea, low
sleep efficiency, stress and turmoil leading up to the event,
E.E.G. results, a sleep hypnogram, significant nocturnal
oxygen desaturations, and abrupt arousals from sleep caused
by a lack of oxygen, all information Dr. Kushida did not
have. Arousals from sleep result when the brain realizes that
a person is not breathing. Sleep apnea arousals can produce
sleepwalking. Dr. Erman’s explanation was as follows:
But let me clarify one of the things I am not
sure we actually stated very clearly and that is
the relationship between sleep apnea and the
capacity to provoke arousal events that may
lead to an episode of sleepwalking.
When someone is experiencing a sleep
apnea event, they are suffocating. The airway
is closed off. The oxygen level is falling. The
levels of carbon dioxide in the body are rising,
and it is a stressor that typically will lead to
something very much like a classic fight-or-
flight response. So when the body recognizes
the body and brain recognizes this event is
going on after amorphizing a little bit, but the
event is ended by an arousal. If we don’t
arouse, we die. And that’s why people who
are on sedative medications or alcohol may
die from sleep apnea. This arousal is very
much akin to somebody poking with a stick.
This is the analogy I often use for patients
which explains why people with sleep apnea
are so tired because awakening repetitively
during the night.

-- 20 of 35 --

LIAO V. JUNIOUS 21
The sleep apnea event can be a trigger in
susceptible individuals for these episodes of
sleepwalking. So the relevancy here as well is
that not only would it contribute to the sleep
deprivation that would increase the risk of the
apnea – of the sleepwalking event, it could
also serve as the specific trigger that would
cause the sleepwalking event.
Dr. Erman also testified that Liao’s behavior on the night
of the assault was consistent with not having been a “focused
assault that leads to injuries that might have been expected
had there been an intent to really seriously injure or kill, that
there was amnesia and confusion following the episode.”
Dr. Erman was appropriately skeptical of Liao’s
description of the events. To ensure that Liao was not
manufacturing a self-serving version of what happened on the
night in question, Dr. Erman gave him an occasion to do so,
but Liao stuck to his story:
(Dr. Erman) I actually gave him the
opportunity to embellish on the story because
to satisfy myself that this was an accurate
history that would be consistent with a non-
rem arousal disorder, I wanted to see whether
given the opportunity to embellish the report
to provide with more detail were there
monsters, were there dragons; and he didn’t
provide that, which to me was consistent with
the history and also consistent with his giving
the truthful account and not really providing
me with more gory details even when I gave

-- 21 of 35 --

LIAO V. JUNIOUS 22
him the opportunity and suggested perhaps
those might have been present.
IV
Standard of Review
Although we review de novo a district court’s decision to
grant or to deny a 28 U.S.C. § 2254 habeas petition, our
review of a state court decision is quite deferential. In this
respect, we accept and assiduously apply the Warden’s
statement of our demanding standard of review.
As amended by the Antiterrorism and Effective Death
Penalty Act of 1996 (AEDPA), 28 U.S.C. § 2254(d)
constitutes a “threshold restriction,” Renico v. Lett, 559 U.S.
766, 773 n.1 (2010), on federal habeas corpus relief as to state
prisoners that “bars relitigation of any claim ‘adjudicated on
the merits’ in state court” unless the claim meets one of the
statute’s two exceptions. Harrington v. Richter, 562 U.S. 86,
98 (2011). Under those exceptions, relief may be available if
the state court decision was (1) “contrary to, or involved an
unreasonable application of, clearly established Federal law,
as determined by the Supreme Court of the United States,” or
(2) “based on an unreasonable determination of the facts in
light of the evidence presented in the State court proceeding.”
Id. (quoting 28 U.S.C. § 2254(d)). Only if a petitioner can
survive this threshold review as to claims previously rejected
on their merits by a state court is a federal court permitted to
reach the merits of a petitioner’s claims, reviewing them de
novo. See Panetti v. Quarterman, 551 U.S. 930, 953 (2007)
(“When a state court’s adjudication of a claim is dependent
on an antecedent unreasonable application of federal law, the
requirement set forth in § 2254(d)(1) is satisfied. A federal

-- 22 of 35 --

LIAO V. JUNIOUS 23
court must then resolve the claim without the deference
AEDPA otherwise requires.”); see also Howard v. Clark,
608 F.3d 563, 571–72 (9th Cir. 2010); Frantz v. Hazey,
533 F.3d 724, 735–36 (9th Cir. 2008) (en banc).
A state court decision is “contrary to” federal law only if
it “applies a rule that contradicts the governing law” as set
forth in Supreme Court opinions, or reaches a different
decision from a Supreme Court opinion when confronted with
materially indistinguishable facts. Williams v. Taylor,
529 U.S. 362, 405–06 (2000). A state court engages in an
“unreasonable application” of federal law if it identifies the
correct governing legal principle from the Supreme Court’s
decisions but unreasonably applies it to the facts of the
prisoner’s case. Id. at 413.
The inquiry under 28 U.S.C. § 2254(d)(1) is sharply
circumscribed. First, “clearly established federal law” is
limited to Supreme Court authority that “squarely addresses”
the claim at issue and provides a “clear answer.” Wright v.
Van Patten, 552 U.S. 120, 125–26 (2008); see also Lopez v.
Smith, 135 S. Ct. 1, 5–6 (2014) (per curiam) (grant of habeas
relief reversed where court relied heavily on circuit decisions
and the Supreme Court had failed to address the specific
question presented by that case); Marshall v. Rodgers, 133 S.
Ct. 1446, 1450–51 (2013) (federal habeas court may “look to
circuit precedent to ascertain whether [a federal appellate
court] has already held that the particular point in issue is
clearly established by Supreme Court precedent,” but may not
use lower court authority “to refine or sharpen a general
principle of Supreme Court jurisprudence into a specific legal
rule” or “to determine whether a particular rule of law is so
widely accepted among the Federal Circuits that it would, if
presented to [the Supreme] Court, be accepted as correct”);

-- 23 of 35 --

LIAO V. JUNIOUS 24
Premo v. Moore, 562 U.S. 115, 127–28 (2011); Knowles v.
Mirzayance, 556 U.S. 111, 121–22 (2009); Carey v.
Musladin, 549 U.S. 70, 77 (2006). And, in light of the record
before the state court and the clearly established Supreme
Court precedent, the state court decision must have been
“objectively unreasonable,” and not merely incorrect in the
view of the federal court. Lett, 559 U.S. at 773; Richter,
562 U.S. at 101–02; see also Felkner v. Jackson, 562 U.S.
594, 597–98 (2011) (per curiam). “[E]ven a strong case for
relief does not mean the state court’s contrary conclusion was
unreasonable.” Richter, 562 U.S. at 102.
The standard set forth in § 2254(d) is “difficult to meet
. . . because it was meant to be.” Id.; see also Burt v. Titlow,
134 S. Ct. 10, 15–16 (2013) (“Recognizing the duty and
ability of our state-court colleagues to adjudicate claims of
constitutional wrong, AEDPA erects a formidable barrier to
federal habeas relief for prisoners whose claims have been
adjudicated in state court.”). It “reflects the view that habeas
corpus is a guard against extreme malfunctions in the state
criminal justice systems, not a substitute for ordinary error
correction through appeal.” Richter, 562 U.S. at 102–03. To
that end, it precludes review of any claims previously rejected
on their merits by a state court except in the narrow category
of cases “where there is no possibility fairminded jurists
could disagree that the state court’s decision conflicts with
[the Supreme] Court’s precedents.” Id. at 102. Accordingly,
to overcome the bar of § 2254(d), a petitioner is required to
show at the threshold that “the state court’s ruling on the
claim being presented in federal court was so lacking in
justification that there was an error well understood and
comprehended in existing law beyond any possibility for
fairminded disagreement.” Id. at 103; see also Titlow, 134 S.
Ct. at 16 (“We will not lightly conclude that a State’s

-- 24 of 35 --

LIAO V. JUNIOUS 25
criminal justice system has experienced the ‘extreme
malfunction’ for which federal habeas relief is the remedy.”)
(quoting Richter, 562 U.S. at 102) (alteration omitted);
Johnson v. Williams, 133 S. Ct. 1088, 1091, 1094 (2013)
(standard of § 2254(d) is “difficult to meet” and “sharply
limits the circumstances in which a federal court may issue a
writ of habeas corpus to a state prisoner whose claim was
‘adjudicated on the merits in State court proceedings’”)
(quoting 28 U.S.C. § 2254(d)).
As for whether or not Liao suffered prejudice because of
counsel’s error, our first task therefore is to determine
whether the Superior Court’s application of the prejudice
prong of Strickland as measured under § 2254(d) “was so
lacking in justification that there was an error well understood
and comprehended in existing law beyond any possibility for
fairminded disagreement.” Richter, 562 U.S. 86 at 103.
Specifically, does the Superior Court’s conclusion that there
was not a reasonable probability that the missing evidence
would have resulted in a different result in Liao’s trial survive
this rigorous test?
V
Analysis
We begin our evaluation with the state’s candid admission
at oral argument of the obvious: During the trial, Liao was
unmistakably prejudiced by the absence of both a sleep study
and the additional information sought by Dr. Kushida.
Counsel for the Warden agreed (1) that Dr. Kushida was
“clobbered” on cross-examination, and (2) that Liao thereby
suffered prejudice during the trial in connection with the core
of his defense. We have no doubt whatsoever that this is true.

-- 25 of 35 --

LIAO V. JUNIOUS 26
Dr. Sharma’s rebuttal and the prosecutor’s summation prove
the incontestable validity of counsel’s admission. What was
the fatal flaw according to Dr. Sharma? The lack of a sleep
study. Then, the prosecutor vouched for Dr. Sharma’s
testimony in her summation to the jury, saying,
[Dr. Sharma] was very – I think he was very
professional about that. He could have said,
you know, there is no evidence that the
defendant is a sleepwalker or I don’t think the
defendant is a sleepwalker for X, Y, and Z
reasons. But he tells us there is insufficient
evidence. Why? Because there was no sleep
study that was done.
(Emphasis added).
Why was there insufficient evidence to support Dr. Kushida’s
opinion and therefore Liao’s defense? Because of counsel’s
pre-trial error.
However, counsel describes the sleep study we now have
from Dr. Erman as impeaching Dr. Kushida, not at all
supporting his opinion. From this characterization, counsel
argues that had this sleep study been available during the
trial, it would not have helped Liao – to the contrary.
Therefore, he argues Liao suffered no prejudice.
This argument is unconvincing and patently irreconcilable
with the record. Doctor Erman’s detailed testimony as quoted
earlier speaks for itself. Dr. Erman and Dr. Guilleminault
looked at the objective measurable results of the sleep study
and concluded that Liao was a sleepwalker. Their joint
diagnosis would not only have corroborated Dr. Kushida’s

-- 26 of 35 --

LIAO V. JUNIOUS 27
opinion, but it was predicated on objective criteria that
amounted to direct medical evidence strongly tending to
prove the validity of Liao’s defense. This evidence was not
just corroborative, and certainly more than cumulative. It
was direct essential evidence of Liao’s asserted unconscious
state during the attack. It is inconceivable that the results of
Dr. Erman’s sleep study and testimony would have
impeached Dr. Kushida. Any argument to the contrary is
flatly unreasonable. Moreover, if counsel had Dr. Erman’s
and Dr. Guilleminault’s testimony available at Liao’s trial,
we firmly believe that they would have likely been called as
witnesses before Dr. Kushida took the stand. Why? Because
they had what Dr. Kushida lacked: The objective results of a
sleep study.
The Superior Court’s fact-driven finding that Liao
suffered no prejudice evinces additional consequential errors
which highlight the unreasonableness of its decision.
First, the court failed in its decision to acknowledge Dr.
Erman and Dr. Guilleminault’s sleepwalking diagnosis and
the objective evidence supporting it. Instead, the court
focused out of context on bits and pieces of Dr. Erman’s
testimony which the court regarded as diluting his diagnosis.
In so doing, the court ignored Dr. Erman’s explanation of
why these fragments did not erode his diagnosis.
Second, the court opined that the lay evidence of Liao’s
sleepwalking offered during the trial by his relatives, plus Dr.
Kushida’s belittled opinion, were sufficient to establish
Liao’s defense, making Dr. Erman’s evidence essentially
unnecessary. Equating lay testimony on a medical subject
with the testimony of two qualified doctors makes no sense.
One doubts that there is a lawyer alive who, with doctors

-- 27 of 35 --

LIAO V. JUNIOUS 28
available to prove a medical condition, would use lay
witnesses instead, especially in a criminal trial where a
defendant needs only a reasonable doubt to prevail. Indeed,
the prosecutor pointed out not only the weaknesses in Liao’s
relatives’ testimonies, but that they were biased in his favor.
She told the jury, “These women have biases. But, you know,
what doesn’t have a bias is the evidence.” We agree, the
sleep study also would have had no bias. The Superior Court
may have been impressed with the relatives’ testimonies of
sleepwalking, but the prosecutor certainly was not. Neither
was the jury.
Third, the Superior Court inexplicably brushed off the
penetrating effect of the absence of a sleep study on Dr.
Kushida’s testimony, referencing instead Dr. Kushida’s claim
that a sleep study was optional. The court said, “The fact that
Dr. Kushida was aggressively cross-examined is not a basis
to grant a second opportunity to present its case but is
something that happens in the search for truth.” Liao doesn’t
complain about aggressive cross-examination per se, but that
his counsel’s error made the cross-examination brutally
effective – as Dr. Sharma’s rebuttal testimony undeniably
demonstrates. We note that without a study, Dr. Kushida was
left with only an opinion, not a diagnosis. Dr. Erman came
forth with a diagnosis. Moreover, the Superior Court did not
even mention Dr. Guilleminault, identified as the leading
expert in his field.
Fourth, the Superior Court discredited the sleep study
because it did not produce an episode of actual sleepwalking
on Liao’s part. The court said, “Obviously, if the sleep study
had noted a full episode of sleepwalking, there would be no
question that the outcome probably would have been
different.” This comment and expectation ignored what Dr.

-- 28 of 35 --

LIAO V. JUNIOUS 29
Kushida said in his April 8, 2003 letter to the court about
what the study might show: “However, it is very unlikely to
capture an actual sleepwalking episode by polysomnography
. . . .”
Dr. Erman agreed with Dr Kushida.
(Dr. Erman) [W]e typically don’t expect to
see the full episodes of arousal in patients
with good histories of these arousal disorders
when they’re in the lab. It’s very often as if
part of the brain is functioning to – to keep an
eye out on what is going on in these new
surroundings. When you add into that the
circumstances of someone who’s been
incarcerated who knows guards are outside
the room and who is shackled, the expectation
would be that this would lighten fragment
sleep; and we did see this, the sleep
efficiencies for the two studies we did were
both in the low range and 60% range. We
would ordinarily expect that to be in the high
80's [sic] to low 90's [sic].
The court simply overlooked this evidence and improperly
substituted its understanding of the expected results of a sleep
study for those of qualified doctors.
Fifth, in discussing the facts surrounding the early
morning attack, the court highlighted only those that might
prove Liao was conscious and aware of what he was doing,
omitting those that did not. From these selective facts, the
court concluded that in comparison to the defense, the
prosecutor’s case was strong. What the court overlooked and

-- 29 of 35 --

LIAO V. JUNIOUS 30
did not discuss were those contemporaneous facts offered by
Henry Chen and his mother that suggested Liao was not
aware of what he was doing. As indicated in Part I of this
opinion, those facts are:
1) At midnight, just four hours before the attack, Li’s
sons Henry and Danny got into a fight. Liao helped
Li separate them, and when Li started to punish them,
Liao intervened on their behalf and asked her not to
do so.
2) Within seconds of the attack, when Chen asked Liao
what he was doing, Liao did not reply. Liao appeared
to be in shock.
Q. What do you mean by that, he seemed to
be shocked?
A. (Henry Chen) Like open his mouth and
just looking at me.
Q. Sorry? Open mouth, and what else?
A. (Henry Chen) And staring at me.
3) When Li asked Liao immediately after the attack why
he hit Chen, Liao repeatedly said, “Why did I do it?”
His answer was that he had a dream someone was
hitting him and he was fighting back.
4) Immediately after the attack, Liao assisted Li to tend
to her son’s wounds.
Q. What was [Liao] doing?

-- 30 of 35 --

LIAO V. JUNIOUS 31
A. (Li) Nothing. Walking back and forth.
And also he was calling 911.
. . . .
Q. Was he helping you and your son that
evening?
A. Yes.
Q. And how was he helping you and your
son?
A. He asked me to examine my son’s injury
and to put something over the injury to
stop the bleeding.
Q. Did he seem concerned about your son
and his injury?
. . . .
A. Yes. He was trembling all over at seeing
my son’s bleeding.
5) Liao did not flee, he called the police and
accompanied Li and Chen to the hospital.
In other words, Dr. Erman and Dr. Guilleminault’s diagnosis
would not have existed in a vacuum, but would have served
to explain and to interpret Liao’s behavior that was not
consistent with trying to murder Li’s son with premeditation.

-- 31 of 35 --

LIAO V. JUNIOUS 32
VI
We conclude that the Superior Court’s fact-based decision
that Liao suffered no prejudice from his counsel’s error was
not just merely incorrect, but “objectively unreasonable.”
Lett, 559 U.S. at 773; see also Richter, 562 U.S. at 101–02.
What is more, the Superior Court’s application of Strickland
to the facts of this case was demonstrably unreasonable as
that term has been construed in this context by the United
States Supreme Court.
Accordingly, deference to the state’s decision is not
applicable. Milke v. Ryan, 711 F.3d 998, 1008 (9th Cir.
2013). Thus, we look de novo at this issue. From this
perspective, we note that our precedent recognizes that
prejudice is established when, as Liao’s counsel argues,
“counsel’s error left the defense with weaknesses that were
exploited by the prosecution.”
In Brown v. Myers, 137 F.3d 1154 (9th Cir. 1998), for
example, counsel’s error was his failure to call available
witnesses who could have corroborated his client’s alibi
defense. We said,
The district court concluded that the alibi
witnesses would not have helped Melvin at
trial because their testimony during the
evidentiary hearing was vague with regard to
time. Their testimony, however, was
consistent with Melvin’s account that he
arrived at Saunders’ house too early to have
participated in the shooting. Because their
testimony buttressed Melvin’s account on this
crucial point, it creates a reasonable

-- 32 of 35 --

LIAO V. JUNIOUS 33
probability that the fact-finder would have
entertained a reasonable doubt concerning
guilt. As it was, without any corroborating
witnesses, Melvin’s bare testimony left him
without any effective defense.
Id. at 1157–58 (citations omitted).
We came to a similar conclusion in Luna v. Cambra,
306 F.3d 954 amended in 311 F.3d 928 (9th Cir. 2002),
another case involving a failure of trial counsel to call known
alibi witnesses to corroborate his client’s testimony. Citing
Brown, we determined that counsel’s error prejudiced Luna
because without corroborating witnesses, his “bare testimony
left him without any effective defense.” Luna, 306 F.3d at
961 (quoting Brown, 137 F.3d at 1158).
The validity of our precedents as they relate to this appeal
finds support in a recent Supreme Court case, Hinton v.
Alabama, 134 S. Ct. 1081 (2014). The Court said that a
defendant could well be prejudiced by his attorney’s failure
to secure an expert witness on a scientific issue if “there is a
reasonable probability that . . . [the] expert . . . would have
instilled in the jury a reasonable doubt as to [his client’s] guilt
. . . .” Id. at 1089. The Court could easily have been talking
about Dr. Erman.
The magistrate judge’s decision was similarly flawed.
The decision erroneously labeled Dr. Erman’s testimony as
well as the results of the sleep study “merely cumulative.”
This label betrays a serious misunderstanding of the
difference between direct and corroborating evidence, on one
hand, and evidence that unnecessarily proves a point already
sufficiently established, on the other. Black’s Law Dictionary

-- 33 of 35 --

LIAO V. JUNIOUS 34
defines cumulative evidence as “[a]dditional evidence that
supports a fact established by the existing evidence (esp. that
which does not need further support).” Evidence, Black’s
Law Dictionary (10th ed. 2014). Corroborative evidence, on
the other hand, is “[e]vidence that differs from but
strengthens or confirms what other evidence shows (esp. that
which needs support).” Id. Articulating these definitions
exposes the fatal error in degrading Dr. Erman’s and Dr.
Guilleminault’s input and the sleep study results to
unnecessary surplusage.
The heart of Liao’s defense was lack of intent caused by
a sleep disorder. The Superior Court correctly articulated the
importance of this issue, saying
The evidence in this case centered upon
the issue of consciousness. If Mr. Liao was,
in fact, in an unconscious state, under the law,
he would not be responsible for his act. An
unconsciousness would preclude an express
malice or the intent to kill. It would preclude
deliberation and premeditation because,
obviously, one cannot formulate an intent to
kill or deliberate or premeditated [sic] in an
unconscious state.
Counsel’s error left Liao’s defense weak and pregnable. It
would not have been so with the evidence the jury did not
hear because of counsel’s mistake.
VII
Liao has served his time in prison and is currently on
parole. It is difficult to conceive of circumstances under

-- 34 of 35 --

LIAO V. JUNIOUS 35
which the State would again take him before a jury.
Nevertheless, as is our practice and authority, we reverse the
decision of the District Court and remand with instructions to
grant a conditional writ of habeas corpus ordering Liao’s
release from all forms of custody unless the State of
California elects within 90 days of the issuance of the
mandate to retry him. Any such retrial shall commence
within a reasonable time thereafter to be set by the District
Court.
REVERSED AND REMANDED.

-- 35 of 35 --